1 Introduction

Plasma diagnostics is an integral part in the progress and understanding the formation and behavior of the plasma for fusion programs. A variety of plasma diagnostic techniques have been developed for the study and control of various parameters of the tokamak plasma [1]. Among these diagnostics, Thomson scattering [2–4] is one of the basic and essential diagnostics to get precise and absolute value of plasma temperature and density with high spatial and temporal resolution. Though this technique is direct and non-perturbing, but due to the small electron scattering cross-section and hence a poor signal-to-noise ratio the implementation of Thomson scattering becomes difficult. To improve the signal-to-noise ratio of scattered photons, it is required to have a powerful source of probe radiation/laser. It may also be noted that along its propagation path across the plasma, the input laser beam/probe beam passes through vacuum windows at the entrance and exit interface of the vacuum vessel. During its passage the laser beam gets scattered from these windows and also from in-vessel components and vessel walls [5]. Thus, an increase in input energy of the incident laser beam correspondingly increases the stray light level. In order to reduce the level of stray signal, apart from mandatory precise alignment of laser beam in the system, various methods are frequently employed, e.g. moving the entrance and exit windows far away from the collection area, use of light baffles and termination of the field-of-view of the imaging system on some suitably chosen viewing dump. The major contribution to the stray signal comes from spuriously scattered radiation that enters into the solid angle of the collection optics after reflection from the vessel wall exactly opposite to the collection port. Thus a viewing dump, which can absorb most of the incident light, is required to cover this area. In view of this a proper design and optimization of viewing dumps in vacuum vessel is very important for detection of Thomson scattered signal. Viewing dump also plays a major role in other plasma diagnostic systems based on electromagnetic radiation by means of reducing the effect of plasma background radiation on the actual signal [6, 7]. Without the use of a properly designed viewing dump, the measurements can be severely affected by the multiple-scattering/reflection of probe radiation and plasma background radiation off the vessel walls and in-vessel components used for different purposes [8]. Thus it becomes very important to make an analytical experimental study on the viewing dumps for its optimization to minimize the stray signal level.

In the present paper we report experimental studies on the scattering of probe laser radiation from graphite in the form of plane tiles, V-grooves engraved surfaces and from a stack of razor blades, which are frequently used as viewing dumps in Thomson scattering experiments. Angular distribution of scattered signal for different angles of incidence of probe radiation is measured for these surfaces. Measured experimental results on scattering from a stack of razor blades and the V-grooved graphite are discussed in the framework of boundary diffraction wave theory [9, 10]. According to the theory of diffraction, the signal of scattered light is dependent on the number of scattering elements present in the illuminated area. This is confirmed in our experiments on scattering from a stack of razor blades and the V-grooved graphite. It is also observed that in case of multiply edged dumps the amplitude of scattered signal depends on the direction of orientation of the dump edges. We have investigated this experimentally for two directions of orientation (horizontal and vertical) of the dump edges. It is found that in case of horizontally oriented edges stray signal has minimum in the horizontal plane, where the detection systems of many optical diagnostics are aligned on the tokamak’s radial ports. Experiments were also performed with horizontally polarized and vertically polarized incident light, and it is found that the dump is not sensitive to the direction of polarization of incident beam until groove/edge spacing becomes comparable with the wavelength of incident light.

2 Experimental setup and measurement

A laboratory prototype experiment was designed to perform scattering analysis on different materials for the optimization of viewing dumps to be used in actual Thomson scattering system at a tokamak. Scattering of probe laser beam from various optical/mechanical components, other than the plasma in the tokamak, results in the generation of stray radiation. The experimental setup used for the study of scattering of light from the viewing dump materials is schematically shown in Fig. 1. A collimated beam of He–Ne laser (λ=632.8 nm) of diameter 6.2 mm is used to illuminate the sample (dump material) mounted on a calibrated and marked rotary stage with an accuracy of 1∘. The He–Ne laser is used in our experiments because of its compact size, visible wavelength, sufficient power, easy operation and handling and wavelength close to that of the ruby laser (λ=694.3 nm), which is used as probe laser in Thomson scattering systems. Experiments are also performed using Nd:YAG laser (Quantel, model: YG982S, λ=1064 nm) particularly in order to see the effect of wavelength on the scattered signal for p-polarization and s-polarization states of the incident laser beam. Light scattered from the sample is collected with a plano-convex lens (25.4 mm diameter and 50.2 mm focal length) and detected with a PIN-photodiode. The lens and the photodiode are aligned by reflecting the laser beam from a reflecting surface mounted in the sample holder. Both these components are mounted on a single platform, which is placed on another similarly marked rotary stage. Thus, a complete detection system can be put at any desired angular position in the horizontal plane to get an angular distribution profile of the scattered signal for a chosen angle of incidence. A lens is placed at a distance of 180 mm from the illuminated sample material surface. Signal output from the photodiode is displayed on an oscilloscope for scattered signal measurements. Rotary stage loaded with dump sample is rotated to a pre-decided angle for changing the angle of incident beam with respect to the normal of the sample surface. A similar rotary stage loaded with imaging and detection system is rotated to obtain the angular distribution of the scattered light for a particular angle of incidence. For each angle of incidence i.e. 0∘,15∘,45∘, and 70∘ (reported in this work), scattered radiation is measured at equal steps of 10∘. A pyrolytic isostatically pressed dense graphite material in the form of plane tiles and V-grooved surfaces (dimensions 55 mm×45 mm×14 mm with V-grooves having depth of 11 mm on one side of this material) and a stack of razor blades (200 stainless steel blades of 100 μm thickness properly aligned one-over-other and held between two aluminum plates; dimensions of the surface facing the incident laser beam 20 mm×37 mm) are used as dumps in our experiments. For easy visualization of the geometries of the samples used in this study the schematic diagrams of the V-grooved graphite [Fig. 2(a)] and a stack of razor blades [Fig. 2(b)] as dump materials are shown in Fig. 2. For studying the scattering at plane graphite, the planar side of the graphite sample (side opposite to the engraved grooves) is illuminated with the laser beam and to study effect of scattering from the V-grooved dump same material is rotated by 180∘ along its vertical axis so that incident laser beam illuminates the V-grooves. The spacing between consecutive grooves is 2.7 mm and groove angle is ∼14∘.

Fig. 1
figure 1

Schematic presentation of the experimental setup for measuring distribution of scattered radiation from graphite beam dump

Full size image
Fig. 2
figure 2

Schematic representation of the shape of V-grooves in the graphite dump (a) and stack of razor blades (b)

Full size image

3 Theoretical description

Scattering from the beam dumps can be understood using the theory of diffraction of light. A unit amplitude plane beam of light incident at a point A on the dump sample can be represented as

$$ U(A) = \exp(jkr) $$
(1)

Here the time factor exp (jωt) has not been considered throughout the work as most of the detectors do not respond to oscillations in this frequency range and recorded intensity is averaged over a time period quite longer than the wave period. Let us consider that this wave is incident at a point A on the edge of V-grooved graphite or on the blade edge in the stack of razor blades and then gets diffracted from there. According to the boundary diffraction wave theory the diffracted field at the observation point P is given by [9]

$$ U(P) = U^{g}(P) + U^{d}(P) $$
(2)

where

(3)

and

$$ U^{d}(P) = \frac{1}{4\pi} \int_{\varSigma} \frac{\exp \{ jk( r + s ) \} \cos (n,s)\sin (r,dl)}{rs\{ 1 + \cos (s,r) \}}\,dl $$
(4)

In (4), R is the distance from the source to the point of observation, P; s is the distance between a typical point A on the dump-edge and P, Σ denotes the boundary of the illuminated part of the edge, dl is an infinitesimal element situated in Σ and n is the unit vector outward normal to the plane of edge. Here U g propagates according to the laws of geometrical optics and is known as the geometrical wave, while U d is generated from every point of the illuminated boundary of the edge and is called the boundary diffraction wave. This boundary diffraction wave diverges out in all the directions from the edge [11] and its amplitude has maximum value in the forward direction which goes on decreasing with increase in the angle of diffraction. Dependence of the amplitude of the boundary diffraction wave on the angle of diffraction is given by the obliquity factor cos(n,s)sin(r,dl) dl/[1+cos(s,r)] in the expression for the boundary diffraction wave.

In the case of diffraction of light beam from a system consisting of a number of edges such as beam dumps in the form of V-grooved graphite material and the stack of razor blades, the total amplitude of the diffracted/scattered signal will be the combined effect of all boundary diffraction waves starting from all the illuminated edges. The total amplitude can add coherently or incoherently depending on the coherence properties of incident radiation. In addition to edge diffraction, diffraction can also take place from surfaces of the edges [12]. In case of beam dumps, light striking at the edge-surface (wall surface of the edge) will be trapped inside the dump due to multiple scattering/reflections between the dump edge walls and only a small portion will come out of the V-grooves. The amount of escaped light from the V-grooves depends on the groove depth, groove spacing and orientation of the grooves in addition to the absorption coefficient of the material. The groove depth and groove spacing determine the number of scatterings/reflections Q, taking place inside the groove walls by the relation: Q=π/θ, for a ray normally incident on the dump [8] where θ is the groove angle. A fractional value of Q implies that the emergent beam is not normal to the dump. The effect of the direction of orientation of the grooves on the scattered signal has been studied experimentally and results are reported in the next section. It may be noted that for diffraction of light from a system consisting of a number of edges in an optically opaque material, only scattered/reflected light (boundary diffraction wave) is available for detection as no directly transmitted or specularly reflected geometrical light is produced. Consequently only the boundary diffraction wave U d is responsible for the amount of stray signal reaching the detector and the term corresponding to geometrical wave can be dropped from the expression for total diffracted field [Eq. (2)]. Thus, for the case of diffraction of a laser beam from N equi-spaced edges, total amplitude of diffracted light U D(P) is given by coherent superposition (considering laser has sufficient coherence properties) of all the boundary diffraction waves starting from N edges and may be represented as

(5)

The phase difference between successive boundary diffraction waves corresponding to an optical path difference Δ is given by δ=(2π/λ)Δ, where λ denotes the wavelength of incident radiation. It may be noted that for effective selection of edge separation that results in minimum of the scattered light for a particular wavelength (wavelength of probe laser), the scattered boundary diffraction waves must satisfy the condition of destructive interference at the detection system. This condition can result in complete destructive interference only if the phase difference for various boundary diffraction waves is 180∘ and their amplitudes are equal.

4 Results and discussion

4.1 Angular distribution of scattered radiation for different angles of incidence

The angular distribution of scattered signal for a p-polarized He–Ne laser beam incident at different angles of incidence 0∘, −1∘, −45∘ and −70∘ on the surface of plane graphite, V-grooved graphite and a stack of razor blades is shown in Fig. 3. The V-grooves in the graphite material and in the razor blades stack are oriented in the horizontal direction i.e. parallel to the plane of incidence. Two orientations of edges/V-grooves named horizontal direction [parallel to the plane of incidence, Fig. 4(a)] and vertical direction [perpendicular to the plane of incidence, Fig. 4(b)] used in our studies are schematically shown in Fig. 4. Here it can be noted that the angles measured on the opposite side of the surface normal relative to angle of incidence are taken as positive while angles measured on the same side of the surface normal relative to the angle of incidence are taken as negative. Same convention is used for the angles of incidence and for the angles of scattered radiation. All the results reported in the paper are normalized reflectivities of the respective dumps. The normalization is performed by dividing the reflected/scattered signal per unit solid angle with the signal corresponding to the incident light intensity. The results reported in Fig. 2 depict that the amplitude of the scattered signal is dependent on the angle of the incidence as well as on the geometry/surface condition of the material sample.

Fig. 3
figure 3

Angular distribution of scattered signal for different viewing dumps for angles of incidence i=0∘, 15∘, 45∘ and 70∘ (a) scattering at plane graphite surface, (b) scattering at V-grooved graphite dump and (c) scattering at a stack of razor blades

Full size image
Fig. 4
figure 4

Schematic diagram showing the direction of orientation of V-grooves/edges of the dumps in (a) horizontal orientation i.e. V-grooves oriented parallel to the plane of incidence and (b) vertical orientation i.e. V-grooves oriented perpendicular to the plane of incidence

Full size image

4.1.1 Scattering at plane graphite

In the case of scattering from plane graphite surface (Fig. 3a), measured scattered signal for angles of incidence of 0∘ and −15∘ shows a peak centered around their respective directions of geometrical reflection. Here amplitude and angular distribution for both the cases are in the same fashion. For angles of incidence −45∘ and −70∘, the behavior of scattered signal is different from the case of 0∘ and −15∘. The reflectivity profiles for these angles of incidence show an almost plateau type region for a range of scattering angles (−40∘ to 25∘ for i=−45∘ and −60∘ to 50∘ for i=−70∘) and after this it results in a sharp peak. Intensity of the peak becomes very large for incidence angle −70∘, which could be due to specularly reflected light and must be prevented from directly illuminating the detection system.

4.1.2 Scattering at V-grooved graphite

For the V-grooved graphite, the profile of the scattered signal (Fig. 3b) is changed significantly as compared to the scattering from the plane graphite surface. Here, amplitude of the scattered signal is reduced significantly (by 10 times, peak-to-peak intensity) in comparison to the plane graphite surface. For this case also, the scattered signal for angles of incidence 0∘, −15∘ and −45∘ shows the same trend in terms of maximum intensity and angular distribution. The observed peak intensity of scattered signal is found to be symmetric with respect to the direction of geometrically backward reflection for respective angles of incidence. At an angle of incidence i=−70∘, the peak intensity is higher than other three angles of incidence and two peaks are observed in the scattered signal profile. One peak corresponds to the direction of geometrically backward reflection while the second peak is due to the light scattered in the direction of geometrically forward reflection. For angles of incidence i=−70∘ or more it was observed that in the forward direction light was specularly reflected from the surfaces of the V-edges, as spacing between two edges is quite large as compared to the wavelength of the incident light. Here it may be noted that in case of angles of incidence at −15∘ and −45∘, the peak in the scattered signal is observed in the direction of backward reflection instead of in the direction of forward reflection. This is due to the fact that a large amount of light is incident on the edge-surface as compared to the light falling on the edge itself. Amplitude of the scattered signal will depend on the actual illuminated area of the dump seen by the detector. In the geometry of V-grooved dump light scattered/reflected in the forward direction from edge-surface is blocked by opaque base/substrate of the dump material while light scattered/reflected in the backward direction reaches directly to the detector. Hence the peaks are formed in the direction of geometrically backward reflection. For an angle of incidence −70∘, the light reflected in the forward direction is also not blocked by the dump substrate/base and it directly reaches the detector resulting in the peak formation in the direction of the geometrically reflection as seen in Fig. 3(b).

4.1.3 Scattering at a stack of razor blades

The angular distribution profile of scattered radiation from the stack of razor blades is shown in Fig. 3(c). At incident angle of 0∘ and −15∘, the scattered signal is suppressed in comparison to plane and V-grooved graphite. It may also be noted that the peak of the scattered signal is not as sharp as in cases of plane and V-grooved graphite, and the scattered signal is almost uniformly distributed for scattering angles of −90∘ to 90∘. For other two angles of incidence i.e. −45∘ and −70∘, scattered signal is quite less as compared to the signal from plane and V-grooved graphite but the angular distribution shows two intensity peaks symmetrically displaced from the surface normal. For i=−45∘ these two peaks have almost the same intensity and the same angular dependence while for i=−70∘ peak intensity in the direction of geometrically forward reflection rises sharply. These two intensity peaks are formed in the direction of geometrically forward reflected and the geometrically backward reflected beams. For i=−45∘ both peaks have comparable intensity due to large number of scattering edges per unit illuminated area and small spacing between the blades. Here, light scattered in forward direction comes mainly from multiple edges of dump blades in accordance with (5), while the peak formed in the direction of backward scattering has contribution from edge-surfaces in addition to the scattering from multiple edges. It is known from the theory of diffraction that diffracted light has maximum amplitude in the forward direction (obliquity factor dependence of the amplitude of diffracted fields [9]) as is also confirmed by a sharp peak in the direction of geometrical reflection for i=−75∘.

4.2 Effect of sample surface geometry on scattered radiation

To compare the effect of sample surface geometry on scattered signal, the intensities of the scattered signal from plane graphite, V-grooved graphite and a stack of razor blades are together shown in Fig. 5 for angle of incidence i=0∘. These results show that the amplitude of the scattered signal from the plane graphite surface is much higher than that from V-grooved graphite surface and from the stack of razor blades whereas light scattered from razor blades has minimum amplitude. For plane graphite surface scattered signal has a sharp peak around the direction of geometrical reflection, while in case of razor blades the scattered signal does not show such a sharp peak and is almost uniformly scattered over a large angular spread. The peak signal from V-grooved graphite is much weaker than peak signal from plane graphite but is still large in comparison to the signal from the stack of razor blades. It is known from the theory of diffraction that total amplitude of scattered light depends on the number of scattering points and hence on the area illuminated by the incident beam. In case of multiple edges engraved on the surface, the effective surface area of the dump increases with an increase in the number of edges. Thus, according to theory, the amplitude of scattered signal will increase from edged dumps as compared to that from plane surfaces. But in actual practice with increase in number of edges the amplitude of scattered signal decreases until the spacing between grooves becomes comparable to the wavelength of incident light. This is due to the fact that increase in the number of edges divide the surface area of substrate in line area of the edges and area of edge-surfaces forming V-grooves. This will result in multiple reflections of the beam incident on edge-surface, thereby decreasing the contribution of the edge-surface scattered signal considerably and a major contribution to scattered signal comes from edge diffracted/scattered waves only. Thus, in the case of edged surfaces, though the overall surface area of the dump is increased the effective surface area seen by the detector is reduced considerably, which results in the reduction of the experimentally observed scattered signal. The effect of the number of edges present in the illuminated area can be seen in the results presented in Fig. 5. The observed scattering from stack of razor blades is less than that from V-grooved graphite even though stainless steel has large normalized reflectivity as compared to graphite [5].

Fig. 5
figure 5

Comparative angular distribution of scattered signal for different viewing dumps for angles of incidence i=0∘

Full size image

4.3 Effect of edge orientation on scattered signal

The effect of orientations of the edges of the dump V-grooves on the amplitude of scattered signal was also studied as with change in orientation of the edges the effective edge-surface area seen by the detector also changes. To study this, measurements are made on horizontally oriented (parallel to plane of incidence) and vertically oriented (perpendicular to the plane of incidence) edges of V-grooved graphite and the stack of razor blades. The experimental results for these cases for a common angle of incidence of −45∘ are shown in Fig. 6. In case of V-grooved graphite scattered signal from V-grooves oriented in vertical direction is larger compared to V-grooves oriented in horizontal direction. The difference in scattered signal for these two orientations of graphite V-grooves is not very much but is quite large in case of a stack of razor blades. This large difference between the two cases is due to the presence of a different number of diffracting elements in unit area of the laser beam. In case of V-grooved graphite, groove spacing is quite large and only few edges of the graphite were illuminated by incident laser beam, while for the case of razor blades number of illuminated edges is quite large as compared to graphite edges. Further difference in the intensities of the scattered signals for two different orientations of the edges can be explained using the theory of diffraction of light. It is well known from the theory and experiments on diffraction of light from knife-edge that a diffracted signal has maximum value in Keller’s cone [12]. It states that diffracted light has large signal in the cone defined by the incident geometrical beam and is formed in a direction transverse to the orientation of the edge. Thus, in case of vertically oriented edges, the maximum of diffracted light will be in horizontal plane while for horizontally oriented edges Keller’s cone is formed in the vertical plane containing the edge. Also light diffracted from multiple edges reinforces the total signal.

Fig. 6
figure 6

Experimental results depicting the effect of V-grooves/blades-edges orientation on the scattered signal

Full size image

4.4 Effect of polarization of incident light on scattered signal

The studies on V-grooved graphite and the stack of blades have shown that by increasing the number of edges, the amplitude of scattered signal goes on decreasing, but one cannot go on increasing the number of edges infinitely because if edge spacing becomes less than the wavelength of incident light, two new phenomena come into effect. One is the strong enhancement in reflected/scattered signal from the sub-wavelength apertures [13] and the second is that the system becomes sensitive to the state of the polarization of the incident light. It will require a particular orientation of the edges to avoid the direction of the strong scattered radiation. In this case, the amplitude of the scattered signal is given by the well-known Fresnel equations for both horizontally polarized incident beam and for vertically polarized incident beam. In order to observe the effect of the polarization of the incident light on the scattered signal, we used a He–Ne laser in conjunction with a half wave-plate for λ=632.8 nm to rotate the state of polarization of incident beam. Scattered signals for these two states of polarization from plane graphite, V-grooved graphite and a stack of razor blades for an angle of incidence i=−45∘ are shown in Fig. 7. Light scattered from the plane graphite surface shows strong dependence on the state of polarization of incident light. In this case, scattered signal is quite large and sharply peaked for s-polarized incident light as compared to p-polarized incident light. Light scattered from V-grooved graphite does not show any perceptible dependence on polarization. In case of scattering from the stack of razor blades, the signal for s-polarized beam is larger than for p-polarized incident beam. For V-grooved graphite and the stack of razor blades, the angular distribution of the scattered signal for both cases of polarization follows the same trend i.e. presence of peak and angular dependence. In order to see the effect of the state of polarization of the incident light on the scattered signal for the fundamental wavelength of Nd:YAG Laser (λ=1064 nm) studies are made on the same V-grooved graphite dump. A half wave-plate for λ=1064 nm is used to rotate the state of polarization of the incident laser beam. The experimental results for p-polarized and s-polarized laser beam normally incident on the V-grooved graphite dump with edges oriented in horizontal direction i.e. parallel to the plane of incidence are shown in Fig. 8. These results show that in this case also there is no observable effect of state of polarization of incident light on the scattered signal.

Fig. 7
figure 7

Response to the state of polarization of incident light of plane graphite, V-grooved graphite and stack of razor blades

Full size image
Fig. 8
figure 8

Response to the state of polarization of incident Nd: YAG laser beam (λ=1064 nm) of the V-grooved graphite dump with V-grooves oriented in the horizontal direction i.e. parallel to the plane of incidence

Full size image

It may be noted from the above presented results that in case of using these materials and structures as viewing dump in Thomson scattering system in tokamak, certain precautions are required. In case of plane graphite tiles as viewing dumps, the scattered signal for all angles of incidence is quite large as compared to the other two cases of V-grooved graphite and stack of razor blades. Light scattered from stack of razor blades shows minimum amplitude of scattered signal in these studies. A stack of razor blades has been used in past as viewing dump for Thomson scattering system on tokamak [14], however, its use on tokamak has been discontinued due to high heat loads and large magnitude of magnetic field. But a stack of razor blades can be used in systems other than tokamak with smaller or no magnetic field. While using stack of razor blades the main factors are the (i) state of polarization of incident radiation (this is required only when edge separation is comparable to the wavelength of incident light) and (ii) the direction of orientation of the blade edges. For the case of V-grooved graphite dump the amount of scattered signal can be reduced significantly by increasing the number of edges per unit surface area. In practice it becomes difficult to make V-grooves with large depth and having edge separations of the order of the wavelength of incident light in visible spectrum in graphite material with conventional fabrication techniques, but this can be achieved for longer wavelengths of incident radiation. In case of Thomson scattering there are two major sources of radiation: scattering of probe laser beam from in-vessel components and vessel walls and the plasma radiation. Laser scattered signal has a preferred direction of polarization while radiation from plasma is un-polarized. Thus, edges should be oriented in such a manner that radiation from laser has its polarization parallel to the edges of the dump to reduce the level of scattered signal as is evident from results in Fig. 7. The direction of orientation of the edges is also very important, as light scattered from vertically oriented edges is much stronger than from horizontally oriented edges. It is also necessary to take care that no other optical detection system should be placed along the direction of strong scattering.

5 Conclusion

Experimental studies on scattering of laser light from graphite material in the form of plane tiles and V-grooved surfaces and a stack of razor blades are reported in this paper. These materials are frequently used as viewing dump in Thomson scattering systems to reduce the level of stray signal. Present studies are aimed at selection of proper design of viewing dump to reduce the level of stray signal. The angular distribution of scattered signal is measured for different angles of incidence of the laser beam. It is observed that in case of scattering at planar graphite, the distribution of scattered signal has an overall scattered intensity much larger and sharply peaked along the direction of geometrical reflection compared to the other two cases for all angles of incidence. The angular distributions of scattered signal from the stack of razor blades for angles of incidence 0∘ and 15∘ are about two orders less than corresponding scattering at plane graphite surface and more than two times less than V-grooved graphite dump. Further, the profile of scattered signal from the stack of razor blades shows an almost uniformly scattered signal in all the directions without resulting in a sharp peak. In all the cases reported in this study it is observed that for larger angles of incidence, the scattered radiation shows a very large signal in the direction of geometrical reflection, which is in agreement with the theory of grazing reflection. In case of V-grooves/blade-edges the scattered signal forms two peaks for angles of incidence 45∘ or more. The formation of intensity peaks in the profiles of scattered signal is discussed in the light of the theory of diffraction as forward diffracted signal and backward diffracted signal. Measurements on scattered signal from horizontally and vertically oriented dump edges show the edge orientation dependence behavior of the scattered signal. It is found that in case of horizontally oriented edges stray signal has minimum in the horizontal plane. Thus, horizontal orientation of viewing dump edges should be preferred. These observations suggest that use of a properly fabricated V-grooved viewing dump with horizontally orientated edges could suppress the stray signal level more than two orders of magnitude as compared to that from plane graphite tiles as viewing dump and thereby could significantly improve the signal-to-noise ratio in Thomson scattering diagnostic system. The effect of polarization of incident beam on the scattered signal is also measured for He–Ne (632.8 nm) as well as Nd:YAG (1064 nm) lasers. There is no significant effect of polarization for V-grooved graphite having groove spacing of 2.7 mm for both wavelengths.