Andor HoloSpec

The high throughput spectrograph

The Andor HoloSpec spectrograph offers very high throughput, excellent imaging quality, low background scatter and very good spatial and spectral resolution. The present document outlines the key technical aspects of the HoloSpec’s Volume Phase Holographic (VPH) transmission gratings on-axis design over traditional Czerny-Turner based configurations, with a particular focus on the collection power / throughput advantage and high density multi-track capabilities of the instrument.

Introduction

Developing increasingly sensitive spectroscopy detection systems to address the needs of the academic and industrial communities is a continuous challenge. Gathering and detecting more photons ... faster requires not only very sensitive detectors, but also highly optimised optical systems with maximum light gathering power, maximum transmission efficiency, very low background (stray light) and very good spectral resolution and bandpass. The HoloSpec -,in combination with Andor CCD, ICCD and EMCCD spectroscopy detectors - overcomes these hurdles, while also offering a modular, yet rugged and compact platform, with easily exchangeable gratings for maximum flexibility in terms of resolution and bandpass.

This technical note will focus on the following key features of the HoloSpec:

  • High throughput (F/# = 1.8 aperture) – maximum light collection power and transmission
  • High density multi-track spectroscopy properties
  • Low stray light – for maximum signal-to-noise ratio

1. HoloSpec spectrograph overview

The optical layout of the HoloSpec is shown in the schematic of Figure 1. It boasts a compact ‘90°’ folded geometry, and consists of the entrance port where the light is coupled into the spectrograph either through a narrow slit or via an optical fiber. A multi-element collimating lens with optimised AR coatings delivers the collimated beam on to the VPH grating where the light is dispersed and delivered on to the multi-element focusing lens, also designed with optimised AR coatings. The latter focuses the light on to detector at the exit port. An on-axis design is used in the HoloSpec.

Figure 1: View of the internal optical arrangement of the HoloSpec. Internal layout of the HoloSpec is shown on the right.
Key HoloSpec Features
High collection efficiency with ultrafast F/1.8 aperture Acceptance angle of ~32 degrees
6.5x better light gathering efficiency than 320 mm Czerny-Turner spectrograph
100 % light collection from NA=0.22 fiber optics
High throughput optical design High quality AR-coated multi-element lenses – optimised for maximum efficiency in the visible (VIS) or near-infrared (NIR)
High transmission volume phase holographic (VPH) with peak efficiencies as high as 90% (greater than70%)
On-axis imaging-corrected design Minimizes greatly the influence of aberrations compared to off-axis Czerny-Turner designs
Sharp focus in the spectral and spatial dimensions across a wide wavelength range
High signal confinement, allowing very high-density multi-track (multi-fiber) acquisitions with extremely low crosstalk
Gathers more photons per pixel – increased signal-to-noise ratio
Low scattered light Minimizes greatly the influence of aberrations compared to off-axis Czerny-Turner designs
Sharp focus in the spectral and spatial dimensions across a wide wavelength range
High signal confinement, allowing very high-density multi-track (multi-fiber) acquisitions with extremely low crosstalk
Gathers more photons per pixel – increased signal-to-noise ratio
Compact and rugged design Small footprint (L x W x H) 44 x 19 x 17 cm
Excellent thermal stability and easily transportable
Easily interchangeable accessories ‘Snap-in’ accessories, including precision slits, SMA / FC fiber optics adapters and pre-aligned grating assemblies
Slit options from 25 μm to 4 mm (8 mm high)
Specialised Raman grating options Optimized for Stokes/Anti-Stokes, Low-frequency or High frequency Stokes operation, 514.5 to 830 nm laser options
Optional integrated Rayleigh filtering unit Fully-enclosed SuperNotch Plus Kaiser filter compartment with user-friendly external adjustment
Versatile detector mount Output interface compatible with all Andor spectroscopy detectors, userfriendly external focus and detector tilt adjustment.

Optical shuttering can be accommodated with a standalone external shutter placed in the optical path before the entrance port of the spectrograph. The use of a shutter is strongly recommended for multi- track spectroscopy. The HoloSpec base unit is offered in two different platforms where AR coatings are used to optimize the throughput in two distinct broadband regions – one the visible and the other NIR, see Table 1.

Parameter HoloSpec-F/1.8-VIS HoloSpec-F/1.8-NIR
Optimized wavelength range (nm) 450 - 730 800 - 1060
F/# F/1.8 (across entire focal plane) F/1.4 (at center)
Focal lengths Output/Input (mm) 85/75 85/75
Magnification 1.13 1.13

Table 1: Base unit options

Volume Phase Holographic (VPH) grating technology and low stray light

VPH gratings are transmissive optical components and contrast with the standard reflective gratings found in traditional Czerny-Turner (CZ) spectrographs, though they operate on the same principle for the dispersion of the light into its different constituent wavelengths or ‘colours’. A number of new spectrograph systems designed for applications such as Raman, fluorescence or diffuse reflectance-based hyperspectral imaging, micros-spectroscopy mapping, microfluidics analysis, on-line process analysis are now based around VPH technology.

The functioning of the VPH grating compared with the reflective surface relief grating as used in most Czerny-Turner (CZ) systems is illustrated in Figure 2. Holographic techniques are used to form a fringe pattern within photoresist which results in an undulating periodic variation in the refractive index of the material. It is this high frequency periodic pattern which causes the diffraction or scatter of the light into the characteristic dispersed spectrum. The reproducible clean sinusoidal undulation of the refractive index variation in VPH gratings ensures very low scattered light is produced from these gratings.

Figure 2: Principle of operation of a VPH grating contrasted with that of the traditional surface relief reflective grating.

VPH gratings are protected by a glass enclosure, which is the primary reason for VPH-based spectrographs such as the HoloSpec to operate at wavelengths above 350 nm and across the visible and near-infrared range.

These gratings can give very high diffraction efficiencies with some as high as 90% and typical efficiencies around 70% (Barden et.al. (2000). Their design can be tailored to accommodate a wide range of resolution and bandpass combinations depending on the application – this is achieved through a careful choice of groove density and angle of incidence. The HoloSpec features a range of standard gratings optimised for applications such as broadband luminescence or Raman spectroscopy. The gratings may be categorised as follows but it is important to note that a given grating may be used for a range of different applications and not just Raman; it is the wavelength range and resolution which is of key importance:

  • Broadband gratings – available to cover broadband regions from ~400 nm right up to ~1068 nm.
  • Raman low frequency (LF) gratings - available for specific laser excitation wavelengths (514.5, 532, 632.8, 647, 752, 785, and 830 nm) and are designed to access low frequency shifts (~50 cm-1 to ~2500 cm-1 region).
  • Raman high frequency (HF) gratings - available for specific laser wavelengths (514.5, 532, 632.8, 647, 752, 785, and 830 nm) and are designed to access high frequency shifts (~1800 cm-1 to ~4000 cm-1 region).
  • Raman Stokes and Anti-stokes (SA) gratings - available for the same excitation wavelengths but have been designed to have their central wavelength towards the centre of the focal plane so that both low frequency Stokes and Anti-stokes shifts can be observed (typical Raman shifts in range from ~ -1100 cm-1 to ~ 1600 cm-1).

Throughput and collection efficiency: key enablers for low-light spectroscopy

One of the key features and advantages of the HoloSpec is its very high throughput capability. This is clearly desirable when working with low level signals in order to facilitate the capture of as many photons from the sample as possible.

It is worth at this point distinguishing between two similar terms used to describe the system performance:

  • Collection efficiency
  • Throughput

Collection efficiency refers simply to the capability of the system to collect or accept light signal into itself at the entrance port and can be directly determined from the F/# of the system.

Throughput refers to the comparison of the signal out of the system at the exit port with that collected at the input port and takes into account not only the F/# but also the transfer function of the system. The transfer function will take into account reflectivities of optical surfaces, such as those of lenses and mirrors, and the grating efficiencies. The terms are often used interchangeably.

Collection efficiency (and Throughput) is usually specified in terms of the F-number (F/#) of the system or the numerical aperture (NA). These can be related to the acceptance cone angle of the system (2Θ), where Θ denotes the angle between the outer most ray accepted into the spectrograph and the optical axis.

Figure 3 shows a schematic optical layout illustrating the relationship between F/#, NA and acceptance cone angle for an optical system.

Figure 3: Schematic layout of input to spectrograph systems showing the relationship between F/#, NA and acceptance cone angle – 2Θ. Θ (Theta) is the angle that a ray makes with the optical axis of the system and is half the acceptance cone angle. Deff denotes the effective diameter of the optical system defining the diameter of the collimated beam through the system. ‘f’ denotes the focal length of the collimating or focusing optic – either lens or mirror.

Clearly the longer the focal length relative to the effective diameter of the collimated beam through the system, the higher the F/#, the smaller the NA, and the smaller the acceptance cone angle – 2Θ. The acceptance cone angle in turn can be related to the solid cone of rays that can be captured at the entrance port. The extent of this cone of rays is measured in terms of the solid angle (measured in steradians) subtended by the collimating (or focusing optic) at its focal point. This has been termed the TH factor or ‘throughput factor’ in Table 2. The relationships between these quantities are summarised as:

and numerical aperture is given by:

where n denotes the refractive index of the medium: usually the medium is air and the value of n is taken as 1. Finally the solid angle defining the cone of rays that can be captured is given by:

If one assumes that the light signal is coupled uniformly from all directions into the spectrograph, or that light is coupled uniformly at all angles from an optical fiber within its NA, then comparison of the collection efficiency of the different systems can be made by simply taking the ratio of the TH factors

Parameter CZ 300 CZ 320 CZ 500 HOLO HOLO /CZ 300 RATIO HOLO /CZ 320 RATIO CZ300 /CZ320 RATIO
F/# 4 4.6 6.5 1.8 - - -
NA 0.125 0.109 0.077 0.278 - - -
CONE ANGLE (AIR) 14.4 12.5 8.8 32.3 - - -
TH FACTOR (STER) 0.049 0.037 0.019 0.247 5.0 6.6 1.3
WAVELENGTH REGION (NM) Relative insensitivity (counts)
557/579 698 475 225 2.72E+03 3.9 5.7 1.5

Table 2: A comparison of collection efficiency and throughput for some typical Czerny-Turner spectrographs with the HoloSpec. The factors highlighted in orange show the throughput of HoloSpec is ~4 or more times better than the CZ 300 mm focal length system and ~6 times better than that of the CZ 320 mm focal length system.

Table 2 compares the collection efficiencies and throughputs of three Czerny-Turner systems that use toroidal optics for aberration correction with those of the HoloSpec. The first has focal length of 300 mm (CZ 300), the second has focal length of 320 mm and includes an extra optical surface for further image correction (CZ 320), and the third is a system with focal length of 500 mm (CZ 500).

The top section of the table presents the theoretical comparison showing the collection efficiency based on the specified F/# values of each system; this is indicative of the upper limit for throughput that could be reasonably expected – as it doesn’t take into account other factors in the transfer through the system such as grating efficiencies, mirror/lens surface reflectivities and extra optical surfaces.

The lower section of the table is an empirical comparison where the systems have been optimised for the spectral region in the visible around 500 – 600 nm. The CZ systems each had 1200 g/mm gratings blazed at 500 nm, and the HoloSpec had a VPH grating optimised at 532 nm. The exact same camera and exposure conditions were used for the measurements. Light was delivered from the same stabilised light source for each via an optical fiber of 50 μm core diameter with NA of 0.22. The latter corresponds to a delivery input cone angle to each system of ~25° (or in terms of solid angle 0.154 steradians). The specific spectral region chosen for the comparison encompassed the doublet lines from the Hg spectrum at 577 and 579 nm. Relative intensities of the measured signals are given in background corrected counts in the bottom left hand corner of the table.

It is worth noting that since the acceptance cone angle of the HoloSpec is ~32°, and the output cone angle of the fiber is ~25°, then the HoloSpec is able to accept all of the light exiting from the end of the fiber – unlike the CZ systems.

HoloSpec shows a clear light collection power advantage:

  • 6x better when compared with the 320 mm CZ system
  • 4x better when compared with the 300 mm CZ system

This enhanced throughput capability has major implications for improving the signal to noise ratio (SNR) or system sensitivity when measuring really low level signals. Scenarios where this would act as a key enabler are:

  • Micro-spectroscopic measurements on single cells, or quantum structures - where the signal is intrinsically low,
  • Measurements on living cells or samples sensitive to photo-damage - where it is desirable to keep the excitation energy as low as possible to preserve lifetime of sample,
  • Ultra-fast spectroscopic acquisitions - where minimum exposure times are desirable to ensure the fastest rates, resulting in reduced times to actually capture sufficient signal,
  • Enabling sensitive spectral measurements over extended spectral regions - where for example the grating efficiency or detector quantum efficiency (QE) are starting to fall off – with greater throughput available one can get the same SNR for a lower QE detector/high throughput spectrograph combination compared with a high QE detector/low throughput spectrograph combination.

3. Imaging quality – high resolution, low scatter and minimal crosstalk

The HoloSpec offers excellent imaging quality, a performance characteristic which is extremely important when a large number of tracks are needed in multi-track spectroscopy applications. Figure 4 shows the spectral image captured on a 6.4 mm high sensor for a fiber bundle illuminated with an HgNe line source. The fiber bundle consists of 19 x 100 μm core fibers (cladding 125 μm diameter) organised in a linear array with no spacing. This means there is a spacing of 25 μm (the cladding) between each active 100 μm core channel.

Zoomed images of the fiber cores at different spectral wavelengths across the full width of the detector (25.6 mm) show clearly each individual fiber and that minimal crosstalk is to be expected between channels.

An image of the same fiber bundle illuminated with a broadband (BB) source is shown in Figure 5. This shows the potential for high density multi-track spectroscopy on low height sensors. The integrity of the individual channels is maintained across the full width of the focal plane. Intensity profiles are shown at the top of the figure for transverse sections taken across the tracks at the middle, left hand side and right hand side positions, as indicated by the dashed lines in the lower image. The profiles were taken at the center of the sensor and ~1 mm away from the edges.

Figure 4: Imaging quality across focal plane illustrated with spectral image of a 19 x 100 μm cores fiber bundle array on a Newton DU971 sensor (25.6 mm wide, 6.4 mm high).
Figure 5: Spectral image taken with broadband illumination source using a 19 x 100 μm core fiber bundle with the Newton EM DU971 sensor (25.6 mm x 6.4 mm). Enlarged images of the regions on the left, centre and right hand side are shown immediately above. At the top are intensity profiles across the tracks for positions at 1 mm, 13 mm and 24 mm (indicated by dashed lines in main image).

The profiles give some indication of how closely tracks may be placed together and still have minimal crosstalk between channels. Indeed even in this extreme case with no spacing between fibers, it is possible to define narrow tracks, discarding the overlap regions in between and achieve very low level crosstalk. This latter point will be revisited in a later section of this note.

Optical resolution

Resolution of the spectral lines is defined in terms of the full width half maximum (FWHM) of the line. When using optical fibers, this resolution will be determined principally by the diameter of the core of the fiber. Typically core diameters used are 100 μm, 200 μm and the smallest usually 50 μm. In general the spectral resolution that is achievable will depend on the grating used, the slit width or fiber core diameter and the detector pixel size. Typical values for resolution are summarised in the master table of gratings in the spec sheet and manual, where a detector of width 27.6 mm - pixel size 13.5 μm, and a 50 μm slit or 50 μm core fiber are considered at the input port.

Figure 6 shows a line spectrum captured using a 50 μm core diameter fiber with a HSG-532-LF grating in the HoloSpec, and a Newton EM sensor (16 μm pixels). The measured FWHM was less than 4 pixels (16 μm pixels in Newton) which in turn represented a resolution of the system <0.2 nm or in wavenumbers an average of ~6 cm-1 for this spectral region. The corresponding FWHM in pixels with the same setup (HoloSpec with HSG-532-LF grating), when using a 100 μm core fiber is typically 6 pixels (for 16 μm pixels) and the corresponding spectral resolution is 0.3 nm or in wavenumbers on average ~9 cm-1 in this spectral region.

Figure 6: Spectral profile for HoloSpec and Newton EM DU971 camera where the light from an HgNe spectral line source is coupled into the spectrograph with a 50 μm core fiber. Good resolution is evident across the full width of the focal plane: the FWHM is <4 pixels (16 μm pixels) with a corresponding spectral resolution < 0.2 nm across the full width.

Closer detail of the spectrum is illustrated in Figure 7 where images of the captured Hg doublet lines at 577 and 579 nm are shown and the corresponding vertically binned spectral profiles: each point corresponds to an individual pixel.

Figure 7: Images of the captured data for the spectrum in Figure 6 showing the raw signal images for the Hg lines at 577 and 579 nm and their corresponding vertically binned spectral profiles. The FWHM is <4 pixels (16 μm pixels) with a corresponding spectral resolution < 0.2 nm. The points correspond to the individual pixels.

There are several key points to note from the data displayed in Figures 6 and 7:

  • The quality and confinement of the image for each spectral line,
  • The low scatter evident as can be judged from the background baseline beyond 2 x FWHM of each spectral feature (related to high fidelity sinusoidal refractive index profile of the grating),
  • The excellent resolution across the full width of the focal plane (in this case input from 50 μm fiber but equivalent to 50 μm slit if it were used),
  • The excellent confinement of the captured signal in the spatial dimension

Multi-track spectroscopy – high density tracks on narrow height sensors

The excellent imaging quality of the HoloSpec referred to earlier opens up the possibility for doing high density multitrack spectroscopy on narrow height sensors (heights <4 mm). As can be seen in Figure 5 the integrity of the spectral channels is maintained across the full width of the focal plane. Figure 8 shows 21 channels based on 50 μm core fibers with centre to centre separation of ~125 μm (corresponds to a linear array of fibers at the entrance port of the spectrograph) captured on a 3.3 mm high sensor. To assess the level of crosstalk between these spectral profiles the intensity profiles for several vertical sections across the tracks are shown at the top of Figure 8.

The three vertical intensity profiles correspond to different positions along the sensor at the left hand side (1 mm), the centre (13 mm) and on the right hand side (26 mm). The intensity in between the tracks falls almost to zero indicating very low crosstalk between channels before discarding any pixel rows. By defining the tracks accordingly and discarding one or two pixels rows in the overlap regions one can ensure practically zero crosstalk from one spectral track to the other. This means in effect that:

  • Practically all the useful signal can be captured without crosstalk
  • Resolution is achieved which is comparable to much longer focal length Czerny-Turner systems
  • Resulting in excellent resolution and excellent throughput in the one system at the same time
Figure 8: Multi-track spectra on a 3.3 mm high sensor area where each track corresponds to the spectral signal from a 50 μm core diameter fiber delivering light from a broadband (BB) source. It was possible to fit 21 channels on to the 3.3 mm high sensor. Zoomed images show the tracks in more detail at the left, centre and right hand side positions. Intensity profiles at three different positions indicated by the dashed lines in the bottom image are overlaid and shown at the top.

Curvature in spatial dimension

On close inspection of Figure 4 it can be discerned that there is the onset of a little ‘curvature’ in the spatial dimension as one moves further from the optical axis of the system. When using regions on higher height sensors (e.g. 13 mm high sensors) that are relatively far from the optical axis this curvature has to be taken into account and handled appropriately. Thankfully this can be corrected for relatively easy in a number of different ways. For low height spectroscopy sensors this is not needed. One can assess the influence of any curvature by full vertical binning (FVB) of the spectrum. Consistent with the image of Figure 4, the influence of any curvature is minimal, i.e. sub-pixel in terms of the variation of the full width half maximum (FWHM) of the spectral profile, for spectral images up to 4 mm in height. Hence there will be an insignificant effect when using FVB for 3.3 mm high sensors.

4. Applications areas where the HoloSpec is a key enabler

  • Intrinsically photon-starved experiments...e.g. Quantum dot photoluminescence, micro-Raman of biosamples, micro-photoluminescence of carbon nanostructures, plasmonics spectroscopy of light harvesting complex or organic light-emitting diode (OLEDs), cathodoluminescence, stand-off chemical detection.
  • When acquisition time is a constraint...Gather enough photons in short periods of time while accessing meaningful signal-to-noise ratio. e.g. micro-spectroscopy chemical mapping, micro-fluidics such as spectrally-resolved flow cytometry, online process control.
  • Minimizing photo damage of photo-sensitive samples...Protect samples from photodegradation and phototoxicity – achieve meaningful signal-to-noise ratio in shorter timescales to minimize overexposure to excitation sources e.g. biomaterials such as live cells or luminescent biotags.

References

HoloSpec manual – (see MyAndor)
Technical note – F/# matching – (see MyAndor)
Product note – The Andor HoloSpec: Configurations and Optimisation
Barden SC, Arns JA, Colburn WS, and Williams JB, ‘Volume-Phase Holographic Gratings and the Efficiency of Three Simple Volume-Phase Holographic Gratings’, Astronomical Society of the Pacific, 112, pp809-820 (2000)

Sign up for the Andor Newsletter!

Receive articles like this one, product launches, press releases and more with our quarterly newsletter focusing on either Physical Science or Life Science. It's free to subscribe and you can opt out at any time.

NAME EMAIL

Physical ScienceLife ScienceSUBMIT