Intense pulsed light devices: Are claimed output parameters valid?

Presented at the 16th European Academy of Dermatology and Venerology (EADV) in Vienna, 16th 20th May 2007

Poster Authors:

Dr. Mario Trelles,

Instituto Mdico Vilafortuny, Cambrils, Spain

Godfrey Town,
Laser Protection Adviser, Haywards Heath, UK

C. Ash,
Optoelectronics engineer, University of Wales Swansea, School of Physical Sciences, Swansea, UK
Correspondence to: Mario A Trelles MD PhD, Instituto Mdico Vilafortuny /ANTONI DE GIMBERNAT FOUNDATION Av. Vilafortuny 31, E-43850 Cambrils, Spain. Tel: +34 977 361320 Fax: +34 977 791024 E-mail: imv@laser-spain.com

Abstract
Background:
High quality IPLs can offer simple, safe and effective treatments for long-term hair removal, removal of benign vascular and pigmented skin abnormalities, skin rejuvenation and acne treatments. However, significant differences in clinical outcome have been recorded among different IPLs despite apparently identical settings. This study investigates the differences in optical outputs of eighteen IPLs in daily clinical use in the UK and proposes a correlation between light-tissue interaction and spectral output as measured by time-resolved photo-spectrometry. Most free discharge systems are unable to generate true long pulse durations to match the thermal relaxation time (TRT) of the target structure (eg hair follicle or blood vessel). The TRT of a typical hair follicle is in the order of 25 55 ms and therefore, for optimum thermal destruction of the hair, the light energy should be delivered in the same time regimen [4,5,6]. Most conventional free discharge IPLs therefore deliver a train of shorter, high energy sub-pulses with variable on and off times to generate an overall duration and average energy density in the range of the TRT of the target chromophore. Whilst this method may produce a clinically measurable result, it is by no means the most efficient way of achieving long term stable hair reduction or long term vessel clearance with the minimal number of treatments and low incidence of adverse effects.

Methods:
Seven key parameters were identified as having an important role to play in the way light interacts with the skin, and therefore have an important function in patient safety and effective treatment. Simple methods were devised to measure energy density, pulse duration, pulse profile, homogeneity of treatment spot, averaged spectral output, time resolved spectral data outputs and physical dimensions of transmission block. These evaluations were made on site in working clinics between patient appointments.

Results:
Using the assessment methods in this study, fluence values in some IPLs were found to be outside of the tolerances allowed for Class IV medical lasers. Several IPLs, which claimed matching of pulse durations to the thermal relaxation times of specific follicular or vascular targets were measured to have pulse durations that were significantly shorter or longer than claimed. Some IPLs claiming square pulse characteristics failed to show constant spectral output across the duration of the pulse or sub-pulses. Poor spot homogeneity was noted in some systems which could lead to significant side effects at higher output levels.

Figure 1Oscilloscope trace of a rising-falling xenon lamp discharge slope, typical of a free discharge IPL (IPL B).

Figure 2Oscilloscope trace of a square pulse xenon lamp discharge profile of a constant current IPL (IPL E).

Conclusions:
This study provides a method to determine accurately all relevant parameters of the emitted light pulses from IPLs. In particular, the spectral footprints measured using time-resolved spectroscopy may have important clinical consequences in terms of anticipating different light-tissue interactions and hence clinical efficacy and safety and identify potential for side effects.

A constant current (square pulse) system generates energy in the same way as a free discharge system by charging a large capacitor or a number of capacitors in parallel, the difference between the two systems is the way in which the energy is delivered to the flashlamp. Constant current systems create a square pulse discharge profile through the xenon flashlamp (Figure 2) at an optimum energy level and this is repeated without variation. As a result, little energy is wasted and the constant current system can operate efficiently and effectively at lower energy levels producing true long single pulses or groups of square sub-pulses. The duration of the discharged pulse or sub pulses of intense light may be measured using a reversed biased photodiode, acting as a light dependant switch (Figure 3). This method records total xenon lamp output pulse duration recorded by the photodiode including all of the optical energy in the decay slope. Most manufacturers represent the spectral output of their IPL as an average spectral measurement taken across the entire pulse duration as measured above. It is useful to note the approximate distribution of energy across the different wavelengths from a xenon lamp characterised by the bulk of the energy in the range 530 nm 800 nm with several specific atomic lines between 820 nm and 920 nm (Figure 4).

Introduction
Most IPL systems have a number of parameters, which the operator can modify to match the patients Fitzpatrick skin type and treatment selected. These parameters may include: exposure time of intense light (pulse duration or total duration 2 of sub-pulses) in ms, energy density (fluence) in J/cm and choice of cut-off filters to eliminate unwanted wavelengths. Each of these parameters is important to produce the most desirable thermal profile in the melanin of the hair follicle or epidermal pigmented lesions, or in the haemoglobins in the blood vessel, whichever is the major chromophore being targeted. Positive clinical outcomes are widely documented in the literature using a number of well-established IPL devices. The recent increase in the number of IPLs produced worldwide without thorough clinical data to support manufacturers claims of efficacy, safety and reliability raises the question of the suitability of these devices for their intended purpose. IPLs can be categorised into two main types by the method which they use to generate and deliver the energy required for light-based treatments, they are: 1. Free discharge 2. Constant current A free discharge system applies a large electrical charge to a capacitor or a number of capacitors in parallel then discharges the entire stored energy directly though the flashlamp, this discharge profile is characterised by a rising/falling slope (Figure 1). As the current through a xenon flashlamp starts with a low initial value, the light spectrum is shifted predominantly towards the infrared end of the spectrum. As the current increases exponentially, the spectrum shifts towards the blue part of the electromagnetic spectrum, then as the current falls away, the optical spectrum moves back to the red end of the spectrum [1,2,3]. This effect of spectral jitter during the pulse means that the optical output not only varies during the pulse but may also differ significantly from one energy setting to another.

Figure 3Measurement of pulse duration using a reversed biased photodiode acting as a light dependent switch.

Figure 4Standardised Spectral Output Measurement showing IPL E with a sharp cut-off at 530 nm and typical xenon lamp spectral profile: intensity is irrelevant.

Methods
Seven key parameters were identified as having an important role to play in the way light interacts with the skin, and therefore having an important role in patient safety and effective treatment. The authors used previously published methods [7 9] for standardised measurements to quantify the following parameters: energy density (fluence) for various typical pulse patterns over the claimed lifetime of the lamp, to establish whether there is any significant deterioration compared with accepted standards for lasers (which might result in under-treatment); pulse duration, to establish whether TRT for specific chromophores could be matched accurately to the pulse duration; pulse profile, i.e. electrical discharge (oscilloscope trace) to the flashlamp vs. manufacturers claims to determine whether the discharge to the xenon lamp is constant (square) or variable (free discharge); the homogeneity of the treatment spot produced by the IPL glass transmission block, to determine whether any hot spots or low energy areas are produced (that may lead to over- or under-treatment); dimensions of the crystal transmission block contacting the skin surface, to allow accurate calculation of energy density values; averaged spectral output across the entire pulse, to identify cut-off filter accuracy and any unwanted or potentially hazardous (UV) wavelengths; the time-resolved spectral footprint of the intense pulsed light (IPL) across the entire pulse duration to allow detailed examination of the spectral structure of individual IPL pulses and consider potential treatment implications.

Results & Discussion

Calibration
The key optical component requiring calibration in an intense pulsed light device is the applicator head containing the xenon flash lamp assembly. This flash lamp has a finite operating life during which output values deteriorate following decay of parts (deposition of material from the plasma on the inside wall of the lamp envelope, decay of coloured glass cut-off filters, erosion of the optical surfaces of dichroic filters and reflectors, etc.). Therefore, the more shots that are fired, the weaker the lamp becomes and the more variable the output spectrum of light produced. Some more expensive devices include rudimentary built-in energy meters and through software programming allow calibration adjustments to be made automatically to compensate for lamp energy density deterioration. Many devices have i-button control over total available shots in each applicator hand set and upon expiry require the user to either return the entire applicator and umbilical to the supplier for refurbishing / exchange or call out a service technician to change lamps, filters, etc., on site. Several manufacturers have user-replaceable lamp cartridges with a pre-programmed number of shots requiring exchange upon expiry. Some devices had unlimited lamps with only a recommendation when to change the lamp. These systems depend upon preventive maintenance by a factory-trained service engineer routinely checking that the output values remain within manufacturers tolerances. In this study, these IPLs were shown to have the least reliable parameters.

Energy Density
Stated device fluence is assumed by users to be the energy density delivered evenly per cm2 from the whole of the crystal block contact surface to the skin and the fluence values displayed by screen software are believed by operators to be accurate (at least to the nearest whole number). For treatments utilizing selective photothermolysis the light energy is absorbed by chromophores in the skin, such as melanin and oxyhaemoglobin and converted into heat energy. As energy is absorbed, the temperature of the chromophore increases and tissue goes through biological changes. The ideal fluence will raise the temperature of the chromophore to a level that causes damage to the target but does not lead to adverse side effects such as burns or blisters. Even the most simple IPL devices at least allow the user some control over energy density, which makes reproducible measurement very important to ensure consistent output and prevent under or over treatment. Measurements in this study were taken for the most popularly used and the highest IPL settings. Maximum fluence measurements are recorded in Table 1 and fluence results were plotted on graphs for the example eighteen devices against the system displayed fluence (or manufacturers claimed fluence in the user manual if not displayed on the IPL screen) (Figures 5 & 6). By measuring devices in routine daily use, measurements were effectively taken at different stages in the

Time-Resolved Spectrum
Few studies have attempted to document methods for measuring IPL pulse durations, and it appears that no previously published study has examined the time-resolved spectral output of IPLs i.e. across each millisecond of the pulse duration [9,10]. Time-resolved spectroscopy is intended to demonstrate the light energy distribution and stability during a light pulse. These measurements on free discharge vs. constant current systems may be helpful in determining whether there is a potential impact on efficacy of absorption of light by the primary skin chromophore targets of interest. Conventional spectrometers need relatively long sample times, rather like the exposure time on a camera and it is this averaging effect that dampens or eliminates the variations in spectral peaks during the pulse. Time-resolved spectral measurements make it possible to assess variations in spectral composition during the light pulse and hence consider potential treatment implications. The time-resolved spectrum in this study was produced using a fast spectrometer and its counterpart Spectra Suite software (OceanOptics Inc, Dunedin, Florida, USA). This spectrometer and software has the capability of sampling the spectrum of light with a minimum integration time of 1 ms by generating 1000 full spectral scans per second across the entire pulse. Time-resolved spectral data of IPL outputs was captured and stored with an optical resolution of a monochromatic source measured as Full Width Half Maximum (FWHM) resolution of 0.035 nm. This fast spectrometer uses a 2048-element linear silicon CCD-array detector (Sony ILX511, Sony Corp, Tokyo, Japan) to capture data into memory every millisecond interfaced to a standard PC via a USB 2.0 port for later analysis. The spectrometer is externally triggered using a breakout box and because of the relatively fast pulse duration of an IPL system the sampling was taken over an extended time period to ensure capture of the data. The source of the intense white light from the IPL system and the spectrometer optical fibre was separated by a distance of 150 cm to 180 cm to prevent saturation of exposed light upon the CCD array within the spectrometer. During testing suitable Shade 3 IPL protective eyewear was worn by the investigators. IPL spectral outputs were calculated and graphically modeled using MathCAD software. For optimal, uniform visual representation of the three-dimensional images, several specific visual aspects were chosen to show the results including, front view of spectral wavelength, time-resolved spectrum view, 3D flying view of all the dimensions in proportion and a view of the plasma discharge.

Figure 5Standardised Energy Density Measurement showing IPL E whose energy output is well within the accepted tolerance of 20% for Class 4 medical lasers (EN 60825).

FIgure 89 IPL devices out of 18 had applicators that were outside of the standard for medical Class 4 lasers (> 20%) Figure 6Standardised Energy Density Measurement showing IPL F whose output energy is only ca. 25% of the stated energy on the device screen display and which is well outside the accepted tolerance of 20% for Class 4 medical lasers (EN 60825).

manufacturers claimed warranty lifetime of the applicator or lamp/filter assembly, which allowed observation of the degree of deterioration in fluence against claimed values. Using this test method, 30 IPL applicators were measured at maximum fluence of which 11 were more than 20% below and 8 were more than 10% above fluence levels given on the device display or claimed in user manuals, even where brand new lamps were tested (Figure 7). Altogether 9 IPL devices out of 18 had applicators that were outside of the standard for medical Class 4 lasers (> 20%) (Figure 8). The authors considered the accuracy of higher fluence values to be of greatest importance owing to the risk of under- or over-treatment.

This study measured several device (IPLs F and P) where the maximum fluence value displayed was vastly lower than the amount of energy actually delivered to skin and measured fluences were frequently outside the generous tolerance range for lasers (20%) set out in BS EN Standard 60825. The shape of long cylindrical Xenon lamps in many intense pulsed light devices and the presence of a simmer circuit not running along the entire length of the visible discharge tube in the applicator head may suggest an uneven output from the crystal glass transmission block with the greatest concentration of energy produced at the centre. For assurance of continuing reliability of the lamp and filter it is useful if these measurements can be repeated throughout the lamp lifetime. If output drifts significantly from the results recorded with a new lamp then procedures should be in place to replace the lamp or treatment head.

Pulse Duration
Measured pulse and sub-pulse durations using a reversed biased photodiode were recorded as an oscilloscope trace to permit measurement of pulse and sub-pulse durations and intra-pulse delay times (Figure 9). This test also served to validate the number of sub-pulses in a pulse train. Using this method, there was generally a poor correlation between manufacturers claims or system-displayed values and the pulse durations measured. The data measured for six of the eight medical CE-marked IPLs was consistent with displayed values (where given). Only fourteen of 31 pulse duration measurements were within 20% of the manufacturers stated or systemdisplayed values. E.g. IPL B claimed single pulse durations of 5 ms, which were measured at 15 17 ms across all programs and settings.

Study Ref A A1 B C D E F G

Stated Pulse Duration (ms) 3 x 10 ms = 30 14 2 x 2.5 ms = 15 5 5 50 1015 1050 20150 20151 10 20 20 5.5/5.5/5.5 short:5.5/5.5 (2) med:3.6/3.6/3.6 (3) long:3.6/3.6/3.6 (3) 40 34.8 123 15 black 15 blonde 3 5 35 35 10 40 30/40/50

Measured Pulse Duration (ms) 3 x 6 ms = 18 14.5 2 x 3 ms = 15.5 1517 1517 51 no data 1051 ms missing pulses missing pulses 6 6.6 8.7 5.4 6.4 7.0 5.5/5.5 (2) 4.0/4.0/4.0 (3) 4.5/4.5/4.5 (3) 40 37 121 2.2 5 3 5 132 132 24 42 30/40/50

H I

J K L M N

O Q

Figure 730 IPL applicators were measured at maximum fluence of which 11 were more than 20% below and 8 were more than 10% above fluence levels given on the device display or claimed in user manuals, even where brand-new lamps were tested.

In one example IPL program for IPL F, one sub-pulse was Figure 9Only 14 of 29 pulse duration measurements were within 20% found to be entirely missing which correlated with the low of the manufacturers stated or system-displayed values. fluence measured for that program compared with others on the same device. Such discrepancies are clearly unacceptable as they may lead to selecting incorrect fluence values and under- or over-treatment of the patient leading to either ineffective treatment or unwanted side effects (Figure 10).

Lamp Discharge Current: Pulse Profile

Many IPL manufacturers promote a square or controlled pulse shape (meaning the pulse of electrical energy discharged across the xenon lamp) and colourful artists illustrations are often used rather than shown as an actual oscilloscope trace. It was apparent that almost all claims recorded in this study made in manufacturers literature for a square pulse were not reflected in our example oscilloscope measurements as the pulses were usually the typical xenon discharge slope (increasing/decreasing) of a free discharge system.

Spot Homogeneity Measurement

The homogeneity of the distribution of output energy at the surface of the quartz or sapphire transmission block is important as uneven distribution of energy to the patients skin during discharge of the flashlamp will create over- or under-treated areas. This difference in energy distribution usually occurs where energy is focused into the centre of the transmission block and thus the perimeter has a lower energy density. In this case, the skin only receives adequate energy in circular or elliptical hot spots within the glass transmission block surface area, resulting in the operator having to overlap the treatment areas to achieve effective treatment. Uneven energy distribution may also occur if the side walls of the transmission block are highly polished whereby the energy may be focused around the perimeter of the blocks skin contact surface creating an picture framing effect with an area of ineffective energy located in the centre leading to potential blotchiness on the patients skin where dischromia or hair regrowth appears in discreet patches. The homogeneity of the energy distribution is measured in this study by capturing the intense pulsed light discharge 100 times per second using a digital video camera (Logitech Webcam, Fremont CA, USA) and storing the individual frames on a laptop computer for selection and subsequent analysis. The best quality compression ratio of the software used by modern digital cameras is utilized to reduce the amount of data stored, which results in the frame rate being dramatically decreased. As this study measured a flash of intense light with pulse durations in the order of tens of milliseconds, the test apparatus had to have the capability to save as many frames-per-second as possible to provide sufficient images to capture the development of the plasma phase in the Xenon lamp. To generate as many frames-per-second as possible, in this study a screen recorder with MJPEG compression at 100 frames-per-second at its highest quality setting was used, which acted as a secondary data recorder and allowed a high frames-per-second capture without compromising quality of the images. To prevent saturation of the diode array detectors at the rear of the digital camera, a sheet of white, 1.5 mm thick PTFE was used to heavily filter the amount of light being exposed to the camera. From a general assessment of the homogeneity profiles produced by the camera system by the authors, it was only possible to grade profiles as poor, moderate or good. This approximation was capable of exposing major discrepancies (Figure 12) and give a general overview of light distribution but did not quantify results.

Only IPLs E and G exhibited a true single square pulse shape confirming that they used partial discharge capacitor technology although close pulse-stacking in devices A and O effectively achieved the same pulse shape and device D showed a nearly square pulse shape. Only a comparison of time-resolved spectral output will demonstrate whether there is any spectral deterioration across the entire pulse duration when comparing these devices (Figure 11). A constant current through the xenon flashlamp may be critically important in the treatment of certain skin conditions to achieve the desired photo therapeutic effect. Potentially, much of the discharged energy may be wasted due to uneven wavelength distribution of the light energy across the pulse duration.

Figure 10Standardised Lamp Discharge Duration Measurements showing IPL system F which is a triple pulse system with one pulse missing, thus one third of the energy is lost. System display values: T1 3 ms, T2 20 ms, T3 4.5 ms, T4 30 ms, T5 4.5 ms (T3 4.5 ms sub-pulse missing). It is assumed that this is due to an error when writing the system software or when calibrating the microprocessor control system.

IPL E

IPL G

Figure 12Example Standardised Spectral Output Measurement showing (left) IPL F with applicators with cut-off filters at 560 nm and 690 nm demonstrating poor (shallow slope) cut-off profiles and inaccurate cut-off (620 nm red rather than 690 nm as stated in the manufacturers user manual) and (right) IPL I with cut-off filters at 645 nm, 695 nm and 755 nm showing good but inaccurate (steep slope) cut-off profiles compared with manufacturers operators manual: intensity is irrelevant.

Explanation of Spectral Graphs

The spectrum analysis of light emitted from flashlamps contains both discrete line structure and continuum radiation. The continuum radiation is blackbody radiation, characteristic of the temperature of the plasma in the discharge. The sharp peaks in the distribution of the wavelength spectrum are due to specific energy transitions of electrons interacting with the xenon plasma. Figure 13 shows a spectral graph of a typical IPL system, with 60% of the total energy found between 550 700 nm (the key wavelengths for hair depilation and vascular abnormalities) and illustrates the proportional distribution of the wavelengths. The line spectra in the infrared region of the graph are more dominant, where bound transitions between energy levels of atoms and ions dominate and these lines are known as the Paschen Series. The xenon plasma results in the emission of photons with a continuous range of energies; thus, the spectrum from the xenon ions causes a continuous spectrum. This is known as a bound-free transition and was discovered in 1939 by Rupert Wildt [11]. A sharp cut off point usually around 500 to 600 nm can be seen on the graphs. This is due to a dichroic coated filter formed on the rear surface of the treatment transmission glass block or a separately located dichroic or coloured glass absorption filter to remove any wavelengths below the chosen cut-off. Shorter wavelengths in the blue and ultraviolet region of the spectrum are heavily absorbed in melanin may burn the patients skin and are therefore removed. In addition, most

IPL C

IPL A

IPL Q

IPL B

Figure 11Only IPLs E and G exhibited a true single square pulse shape confirming that they used partial discharge capacitor technology although close pulse-stacking in devices A and O effectively achieved the same pulse shape and device D showed a nearly square pulse shape.

10 filters (33% of tested filters) 4 filters (13% of tested filters)

10.05%

10.11%

0.54%

0.45%

0.42%

0.17%

3.58%

2.92%

2.43%

1.71%

1.47%

4.31%

0.86%

2.63%

1.68%

4.62%

9.42%

8.87%

7.71%

5.98%

5.26%

0.12%

0.10%

0.17%

0.38%

0.52%

4.33%

0.7%

6.7%

0.74%

1.0%

0.07%

More blue
Relative intensity

0% UV <1% UV 1-2% UV >2% UV

More red/IR

More red/IR

Constant current

2 filters (7% of tested filters) 14 filters (47% of tested filters)

Wasted energy

Figure 13Averaged spectral distribution graph for a typical xenon lamp spectral profile measuring less than 10% of total light energy above 790 nm and less than 0.5% above 950 nm. The bulk of the energy (90%) is in the range of 530 790 nm.

Figure 14Shorter wavelengths in the ultraviolet region of the spectrum may burn the patients skin or may be hazardous to the operators eyes and are therefore removed. Of the 30 applicators tested, 6 IPLs measured more than 1% and two measured more than 2% of unwanted UV output below 400 nm when cut-off filters were set significantly higher.

Pulse duration (ms)

flashlamp cylindrical envelope walls containing the inert xenon or krypton gas are of cerium- or titanium-doped quartz or silica and have the property of filtering out the most harmful of these wavelengths in the ultra violet region i.e. below 400 nm. Measurements in this study included both an average value for the complete pulse or pulse train and time-resolved spectral measurements produced by the tested IPL applicators. The spectrum graphs of the entire output are useful as they give an indication of accuracy of cut-off filter wavelengths and the presence of unwanted ultraviolet or infrared wavelengths. Of the 30 applicators tested, 7 IPLs measured more than 1% and two measured more than 2% of unwanted UV output below 400 nm when cut-off filters were set significantly higher (Table 2 and Figure 14).
560nm Handpiece 690nm Handpiece

Intensity

Intensity

1100 15

1100 60 40

T ime10 (ms)

5 180

Wav

) ( nm gth el en

T ime (ms)

20

180

) (nm gth elen v Wa

Measurements of 29 applicators showed nineteen (65.5%) with cut-off filters that were inaccurate by more than 20 nm versus the claimed cut-off value given by the manufacturer. Only 10 applicators (34.5%) were within 20 nm of the stated cut-off. (Figure 15) A few systems also use filters to cut off infrared wavelengths as these are absorbed predominantly in tissue water and therefore do not contribute to specific heating of either melanin or haemoglobin. The results show clearly that for free discharge systems, despite pulse durations quoted by the manufacturer, the optical energy is often only concentrated in a narrow time domain of the pulse. This difference between specified pulse duration and effective optical output may explain unwanted side effects or suboptimal treatments. Figure 16 shows both the schematic and the corresponding time-resolved spectral footprint for both IPL C (free discharge) and IPL E (square pulse). The pulse duration of IPL C was measured at 15 17 ms and shows an analogous rising and falling optical output. The useful energy, however, is concentrated into only a few milliseconds. IPL E shows almost perfectly stable spectral emission across the entire 50 ms pulse. Both the atomic peak (825 nm) in the background and the bulk of the energy in the foreground spread almost evenly across the entire pulse duration (peak at 555 nm) with only slight attenuation of optical output towards the end of the pulse can be clearly identified.

Free discharge

Partial discharge

Figure 16(above) Schematic illustration of the difference in the spatial and temporal characteristics of a free discharge and partial discharge pulse to an IPL xenon lamp and (below) time-resolved spectral measurements of example IPL devices: free discharge (IPL C) and partial discharge (IPL E).

Conclusions
Measurement of IPL devices is becoming an important issue but is still at a very early stage after being neglected for years because of commercial pressures and a lack of regulation. However the popularity of IPL as a treatment is growing and there is now a definite need for measurement to help improve both safety and efficacy. The measurements in this paper will give technicians working with IPLs a useful tool for checking output consistency and diagnosing performance issues. Discrepancies can be seen between measured parameters and manufacturer claims but caution must be exercised with this comparison because the techniques described in the paper are subject to on-going development. In particular, traceability to national standards is a pre-requisite to accurate, absolute, as opposed to relative, measurements. On the other hand, comparing different devices is useful as discrepancies highlight the need for legislation to produce standard measurement procedures. This study has determined easily reproducible test methods for key parameters of IPL devices and tested their validity on eighteen IPL systems in daily use. As mentioned, further work is required if measurement results are to be made traceable to national standards. It is also hoped that an improved homogeneity test can be developed to provide objective measurable values to ensure even distribution of energy across skin contact areas. Time-resolved optical spectroscopy is useful for service engineers to evaluate the performance of IPL systems in the field and for manufacturers to construct improved systems for future clinical treatments. Expert clinicians may use graphical representations of pulse parameters including time-resolved spectroscopy as a tool to evaluate likely treatment efficacy and identify potential for side effects. Based upon the limited number of different units and applicators measured in this study, the variability of output data versus manufacturers claims indicates a need for wider studies of device parameters and the introduction of international standards for intense pulsed light devices.

695nm Handpiece 645nm Handpiece 755nm Handpiece

Figure 15Standardised Spectral Output Measurement showing (left) IPL F with applicators with cut-off filters at 560 nm and 690 nm demonstrating poor (shallow slope) cut-off profiles and inaccurate cut-off (620 nm red rather than 690 nm as stated in the manufacturers user manual) and (right) IPL I with cut-off filters at 645 nm, 695 nm and 755 nm showing good but inaccurate (steep slope) cut-off profiles compared with manufacturers operators manual: intensity is irrelevant.

10

11

References
1. Trelles M A. The efficacy of a low energy constant spectrum emission device for hair epilation Poster at European Academy of Dermatology and Venereology 12 16th October 2005. 2. Clement M, Daniel G, Trelles M. Optimising the design of a broad-band light source for the treatment of skin J Cosmetic Laser Ther 2005; 7:177 189 3. Clement M, Kiernan M, Ross Martin GD, Town, G. Preliminary Clinical Outcomes Using iPulse Intense Flash Lamp Technology and the Relevance of Constant Spectral Output with Large Spot Size on Tissue Australasian Journal of Cosmetic Surgery 2006; 1:54 59. 4. Clement M, Omi T. The use of a constant spectrum, uniform temporal profile intense pulsed light source for long term hair removal J Cosmetic Laser Ther 2006; 8:138 145. 5. Anderson R.R, Parish J.A Selective photothermolysis: precise microsurgery by selective absorption of pulsed radiation. Science 1983; 220(4596):524 7 6. Altshuler GB, Anderson RR, Mainstein D Zenzie HH, Smirnov MZ. Extended Theory of selective photothermolysis Laser Surg Med 2001; 29:416 32 7. Town, G. Output Data & Claims made for Intense Pulsed Light Devices: A Comparative Evaluation. Abstract No: 122, Lasers Surg Med, 26th Annual Meeting, Boston, April 5th 9th, 2006. 8. Town G, Ash C. Evaluation of Intense Pulsed Light Devices Using Standized Methodology. Abstract No: 41, Lasers Surg Med, 27th Annual Meeting, Grapevine, USA. April 11th 15th, 2007 9. Town G, Ash C, Eadie E, Mosley H. Measuring key parameters of intense pulsed light (IPL) devices. J Cosmetic Laser Ther 2007; In Press. 10. Clarkson DM. The role of measurement of pulse duration and pulse profile for lasers and intense pulsed light sources. J Med Eng Technol 2004; 28:132 6. 11. Wildt, R., Electron affinity in Astrophysics, Astrophysical Journal 1939; 89:295.