* Present Address: E & C Consultancy, 20 Gladstone Street, Hathern, Leicestershire, U.K.

† To whom requests for reprints should be addressed

There are a wide variety of methods available to the dermatological researcher to determine

changes in the mechanical properties of human skin, in vivo. However, to measure sensitive changes in the mechanical properties of the stratum corneum, there is only a small number of instruments and methods that may be used with confidence. This is principally because the majority of available instruments and methods involve relatively large displacements of the stratum corneum, either parallel or perpendicular to the skin surface (1). Consequently, the tissue underneath the stratum corneum will have an unacceptably large effect on the measurement.

Results of force and displacement measurements of skin are typically displayed as a hysteresis loop (Figure 2). Analysis of the gradient of the loop (force/displacement or displacement/force) yields derivatives of the dynamic spring rate (DSR) usually expressed as g/mm (a measure of the force required to stretch or compress the skin per unit extension), mm/N or µm/g (measures of stretching or compression of the skin in response to a given applied force). Such analysis yields information about the elastic properties of the skin. Analysis of the phase lag between force and displacement responses yields information about the viscous properties of the skin.

After 20 years of experience with the GBE within our laboratories we believe that the principle of the GBE measurement is still the best available for measuring sensitive changes in the mechanical properties of the human stratum corneum in vivo. Subtle though important changes in skin elasticity (dubbed "softness" by Maes et al., (3)) in response to the application of moisturising formulae have been measured, as have changes in skin "tightness" due to surfactant damage. Our experience has, however, also highlighted the draw-backs of employing the original Hargens GBE instrument in a modern laboratory. These are as follows:

1. Importantly, the instrument employs an "open-loop" method of control, i.e. during calibration of the instrument, and subsequently in routine operation, one assumes that an applied current equals a given force. As the GBE is calibrated on one point only (3g), linearity is not guaranteed over the whole measurement range of the GBE. In addition, calibration drift over time is certain.

2. The components needed to run the GBE are bulky and dated (function generator, signal conditioner, storage oscilloscope, compressed gas / air).

3. The probe components are fragile and, in our experience, break easily and require excessive servicing when used routinely (for example, the fine copper wires connecting the armature to the body of the probe).

In recent years, the cost of precision has improved greatly. We can now achieve using conventional technology what was achieved previously through Hargens' (1) considerable ingenuity. We have designed and built a new instrument that retains and builds on all the principles of the original GBE, but contains none of the components. This instrument, designated the Linear Skin Rheometer (LSR) is described in the following sections.

INSTRUMENT HARDWARE AND DESIGN

A schematic diagram of the new instrument is shown in Figure 3. A force-controlled miniature d.c. servo (Maxon 23-12, 0.5W rating, supplied in U.K. by Trident Engineering), gearing and leadscrew now replace the GBE solenoid arrangement, and drive the LSR probe. The original Schaevitz 050 HR LVDT in the GBE has been replaced with a unit of linearity 0.3% (15µm) and sensitivity 0.01% (0.5µm) (Solartron type DF2.5, Schlumberger Industries). The force exerted on the probe is now measured directly by a calibrated load beam (Minigram Beam Load Cell, type MBH50, rated 50g, supplied in UK by RDP Electronics) with an overall accuracy of <20mg. The load beam is mounted vertically within the instrument casing.

INSTRUMENT CONTROL

An IBM (or compatible) PC is used to control the movement of the probe and to log force and displacement data. Both force and displacement are monitored continuously at a rate of 1KHz using a 12 bit ADC plug-in card (National Instruments MIO16). The motor is controlled with an analogue output signal also generated by the PC. The desired force/time cycle, which is normally a single sinusoid, is calculated initially and then stored in memory as a table of values. The actual force applied to the probe is compared with the desired value in the table 1000 times a second. A feedback loop is used to control the motor which moves the load cell in such a way as to minimise any discrepancy. The force applied thus follows the desired force/time cycle extremely closely. The control loop uses an algorithm with proportional and integral terms, whose relative weighting can be varied.

The PC logs all the force and displacement values over a complete measurement cycle, which is usually set at 0.33Hz, thus generating 3000 pairs of points over a 3 second cycle. Two waveform plots are then obtained (Figure 5). Three parameters may be obtained from these curves:

T the phase shift between the two signals

The viscous component of the stratum corneum is often inferred by calculating the area of the ellipse shown in Figure 2. A more rigorous approach is to perform a regression on the original sinusoidal data in order to solve the equations:

where

Having solved for these equations, it is then a straightforward problem to solve the integral over one cycle that represents the area of the ellipse:

The LSR software solves the above equations for both elastic and viscous components of the data. These are subsequently displayed, directly after measurement. Note: units of µm/g (1/DSR, a measure of stretching or compression of the stratum corneum in response to a given applied force) will be used as a convenient expression of skin softness in the rest of this paper.

INSTRUMENT SOFTWARE

The LSR software runs on a standard IBM (or compatible) PC of at least 386 33MHz speed, and is sourced in C programming language. The closed loop control employed by the LSR is achieved as follows. As the LSR control loop is a sampled data system, it is essential that a fast real time clock is generated which triggers the measurement of data samples and updates the control signal output. All IBM PCs have as standard a user interrupt (on interrupt vector 0x1c) called the TIMER TICK, which is available for programmers to use as a regular timing source. This timing signal is generated from the 4.192MHz system clock via an Intel 8253 programmable interval timer. The BIOS presets this timer to its full scale of 65536 and, therefore, the TIMER TICK interrupt is normally 18.188Hz. This is far too slow for a sampled data system.

This problem is overcome by reprogramming the 8253 divider to give the desired frequency during the measurement cycle, in this case 1KHz. In order, however, not to disrupt important internal functions such as monitoring disk drive heads, the timing of interrupt 0x8 (the interrupt number actually triggered by the 8253 output) needs to be restored. This may be achieved by not using interrupt 0x1c, but replacing the BIOS interrupt function at 0x8 with the control programme itself. The original timing is derived within the new interrupt 0x8 by installing a simple counter and calling the BIOS interrupt at the correct interval. In this way, a fast timer is generated that allows data sampling at 1KHz but which does not harm other internal PC operations.

CALIBRATION OF THE LSR

A simple calibration jig has been designed and built that allows rapid, absolute calibration of actual force and displacement (Figure 6). Displacement is measured by a 10µm resolution Digimatic Indicator (Mitutoyo (UK) Ltd, Warwick, UK) traceable to NAMAS calibration standards. Force is measured by a 10g load beam (Maywood Load Beam type 49034, Maywood Instruments Ltd, Basingstoke, UK) with <10mg accuracy, also traceable to NAMAS calibration standards.

The load cell calibration factor is expressed in terms of mV signal per volt per gram measured. The output signal is then amplified through a proprietary amplifier (Maywood Amplifier type D2000, Maywood Instruments Ltd) set to a nominal gain of 325. The amplified signal is then converted via the MIO16 interface card (ADC) such that a full-scale reading of 2048 is equal to an input of 10V. The calibration factor is expressed in terms of ADC input reading that equals 1g. The following values are required:

The load cell calibration value taken from its certificate L

The bridge supply voltage measured with a NAMAS calibrated digital volt meter B

The gain of the amplifier measured with a calibrated digital volt meter G

The correction for the digital volt meter taken from its certificate C

Thus, L = mV per V per g at a nominal voltage of 10V,

Output from the sensor for a 1g load = (L * B * C) / 10,

Output from the amplifier for a 1g load = (L * B * C * G) / 10,

ADC input reading for a 1g load = (L * B * C * G * 2048) / (10 * 10).

The LVDT calibration factor is expressed in terms of mV output per volt per mm displacement. The final slope is expressed in terms of ADC input per mm displacement. The following values are required:

The LVDT calibration value from its certificate L

The excitation voltage measured with a NAMAS calibrated digital volt meter V

The correction factor for the digital volt meter taken from its certificate C

The slope value is derived as follows:

Output voltage for 1mm displacement = (L * V * C)

ADC input reading for 1mm displacement = (L * V * C * 2048) / 10

In practice, these calculations are performed automatically by a simple software programme, allowing rapid and simple calibration of absolute force and displacement.

SKIN MEASUREMENT USING THE LSR

For direct comparison with the GBE reproducibility data obtained by Maes et al. (3), the reproducibility of the LSR was estimated by the same method. 40 consecutive identical measurements were performed on the back of the hand of a female volunteer. Results were analysed to determine the coefficient of variation of the measurement.

40 consecutive measurements on the same subject and same site indicated that the coefficient of variation of the measurement was only 2.9% (Figure 7). This demonstrates very good reproducibility of the measurement technique and compares very favourably with the value of 3% obtained by Maes et al. (3) for the GBE. The variation measured is almost certainly due to movement of the subject during the probe cycle. This has always been the main source of error in these types of sensitive measurements and various means have been employed to minimise subject movement during readings (e.g. use of a pre-cast plaster mould by Maes et al. (3)). However, like Cooper et al. (4), we have found that the use of no restraint is preferable and employ a simple sloping table on which subjects rest their hands.

The results of the study using the Nova DPM 9003 to measure the hydration efficacy of products A and B can be seen in Figure 8. Both products induced significant increases (p<0.05; paired t-test vs untreated control) in apparent stratum corneum hydration (as measured by impedance changes) up to, and including, 6 hours after application. Moreover, product A increased stratum corneum hydration significantly more (p<0.05; paired t-test) than product B at all time-points up to and including 6 hours after application. Water exerts considerable influence on the mechanical properties of the human stratum corneum due to its complex interactions with keratin (7,8). This plasticisation of the stratum corneum, an essentially viscoelastic material, has been described as skin "softening" (2,3). In the case of topical application of a moisturising formula, the extent of this softening effect is directly related to the ability of the product to deliver and maintain increased water concentrations within the stratum corneum. This is usually achieved by the delivery of humectant compounds such as glycerol and / or use of occlusive lipidic films. Indeed, in the case of products A and B, product A might be expected to leverage greater increase in stratum corneum hydration due to its higher glycerol content (4% (w/w) glycerol in A, in contrast to 3% (w/w) in B) and formulation (gel network, in contrast to a simple oil-in-water emulsion in B). For products A and B, therefore, one would expect to be able to measure (i) absolute significant increases in softness for both treatments and (ii) differing relative changes in skin softness for both treatments in accordance with their apparent hydration performance.

Results of the study using the LSR to measure the effects of products A and B on stratum corneum mechanics are presented in Figure 9. Both products induced significant increases (p<0.05; paired t-test vs pre-treatment baseline) in skin softness at all time-points up to and including 6 hours after application. Moreover, product A induced greater increases in skin softness than product B throughout the time-course, significantly so (p<0.05; paired t-test) at 6 hours after application. These results compare favourably with the relative hydration profiles of the two products (Figure 8). The LSR is, thus, able to measure subtle changes in stratum corneum mechanics in response to hydration and to distinguish between the effect of topical application of moisturising products with differing relative hydration performance.

The control system used for the LSR provides the instrument with an inherently more accurate and reliable measurement capability because it employs closed loop (feedback) control. The GBE, in contrast, uses an open loop method of control whereby a predetermined current is applied to the solenoid and assumed to be transformed into the desired force. As no determination of the actual force generated is made at the time of measurement, it is difficult to know with certainty the true force applied to the skin. Whilst open loop techniques can be used successfully in perfectly stable environments, no account can be taken of instantaneous fluctuations in such a system (notably, in this case, subject movement). With the LSR closed loop system, the true force applied to the skin is measured at a rate of 1KHz and corrective action taken within 1ms to restore that measured force to the required value. This system helps ensure that the test sequence is reliable, repeatable and can dynamically adjust for the inevitable variations that occur during in vivo testing. Put another way, because this system allows instantaneous compensation of error resulting from the conversion of an electrical signal to a mechanical force, the need for the friction-free gas-bearing arrangement of the GBE is eliminated. This allows the deployment of a compact, efficient and flexible new instrument.

The authors wish to thank Dr Paul Stevens (Paul Stevens Mechanical Design, Tunbridge Wells, Kent, UK) for his significant skill and expertise in the design of the LSR and also Dr Chris Gummer of Procter & Gamble HABC Ltd for his guidance throughout the development project.

(2) C.W. Hargens, III, "The Gas Bearing Electrodynamometer (GBE) applied to measuring mechanical changes in skin and other tissues", in Bioengineering and the Skin, R. Marks and P.A. Payne, Eds. (MTP Press, Hingham, MA, 1981), pp. 113-122.

(3) D. Maes, J. Short, B.A. Turek, and J.A. Reinstein, In vivo measuring of skin softness using the Gas Bearing Electrodynamometer, Int. J. Cosmet. Sci., 5, 189-200 (1983).

(4) E.R. Cooper, P.J. Missel, D.P. Hannon, and G.B. Albright, Mechanical properties of dry, normal, and glycerol-treated skin as measured by the Gas Bearing Electrodynamometer, J. Soc. Cosmet. Chem., 36, 335-348 (1985).

(5) C.W. Hargens III, "The Gas Bearing Electrodynamometer", in Handbook of non- invasive methods and the skin, J. Serup and G.B.E. Jemec, Eds. (CRC Press, 1995), pp. 353-357.

(7) J.L. Leveque, M. Escoubez, and L. Rassneur, Water-keratin interaction in human stratum corneum, Bioeng. Skin, 3, 227-230 (1987).

(8) J.L. Leveque, "Water-keratin interactions", in Bioengineering of the skin: water and the stratum corneum, P. Elsner, E. Berardesca, and H.I. Maibach, Eds. (CRC Press, 1994), pp. 13-22.