Research Article Vol. 13, No.

4 / 1 Apr 2022 / Biomedical Optics Express 1888

Diamond Raman laser and Yb fiber amplifier for

in vivo multiphoton fluorescence microscopy
S HAUN A. E NGELMANN , 1 A NNIE Z HOU , 1 A HMED M. H ASSAN , 1
M ICHAEL R. W ILLIAMSON , 2 J EREMY W. J ARRETT, 1 E VAN P.
P ERILLO, 1 A LANKRIT TOMAR , 1 D AVID J. S PENCE , 3 T HERESA A.
J ONES , 2 AND A NDREW K. D UNN 1,*
1 Department of Biomedical Engineering, The University of Texas at Austin, 107 W. Dean Keeton, Austin,
TX 78712, USA
2 Institute for Neuroscience, The University of Texas at Austin, 2415 Speedway, Austin, TX 78712, USA
3 MQ Photonics, Department of Physics and Astronomy, Macquarie University, NSW 2109, Australia
* adunn@utexas.edu

Abstract: Here we introduce a fiber amplifier and a diamond Raman laser that output high
powers (6.5 W, 1.3 W) at valuable wavelengths (1060 nm, 1250 nm) for two-photon excitation
of red-shifted fluorophores. These custom excitation sources are both simple to construct and
cost-efficient in comparison to similar custom and commercial alternatives. Furthermore, they
operate at a repetition rate (80 MHz) that allows fast image acquisition using resonant scanners.
With our system we demonstrate compatibility with fast resonant scanning, the ability to acquire
neuronal images, and the capability to image vasculature at deep locations (>1 mm) within the
mouse cerebral cortex.

© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement

1. Introduction
Multiphoton microscopy (MPM) is widely used in neuroscience to visualize neurons and
vasculature in mouse models at depths up to approximately 1 mm in the cortical cortex [1–3].
Since MPM relies on the nonlinear absorption of multiple photons to initiate a fluorescent event,
ultrafast pulsed lasers must be used as excitation sources [4]. Many strategies to push MPM
deeper involve modifying the excitation lasers, where their repetition rates or wavelengths are
often the subjects of concern.
Titanium-doped sapphire (Ti:S) lasers have long been the standard excitation source for
MPM. They typically output femtosecond pulses at an 80 MHz repetition rate with a power and
wavelength (700-1000 nm) that can excite many fluorophores efficiently through two-photon
absorption. Researchers can effectively image neural tissue to 800 µm depths using these
oscillators [5]. To push imaging deeper, many groups have turned to amplifier sources with
lower repetition rates and increased pulse energy to improve excitation efficiency without raising
average power to levels where thermal damage is risked [6]. Theer et al. used a Ti:S-based
regenerative amplifier with a 200 kHz repetition rate to enable imaging to 1 mm in the neocortex
[7]. Optical parametric amplifiers (OPAs) are routinely used to image even deeper regions when
operating at similar low repetition rates (∼250-500 kHz) [8–10]. This is made possible by their
longer wavelength outputs (1100-1900 nm). Photons at longer wavelengths are less susceptible
to scattering as they travel through tissue, so the percentage of photons that reach the focal
point increases as the excitation wavelength lengthens [11,12]. OPAs are advantageous for deep
imaging with their high pulse energies and spectral characteristics, but their low repetition rates
can limit imaging speed.
While deep imaging is certainly helpful, so is fast image acquisition. Many new strategies to
improve imaging speed have been introduced, including ones that make use of acousto-optic
modulators and resonant scanning mirrors [13]. When dramatically increasing the scan speed,

#448978 https://doi.org/10.1364/BOE.448978
Journal © 2022 Received 19 Nov 2021; revised 17 Feb 2022; accepted 22 Feb 2022; published 4 Mar 2022
 Research Article Vol. 13, No. 4 / 1 Apr 2022 / Biomedical Optics Express 1889

it is crucial to ensure the excitation laser’s repetition rate can keep up with the decreased pixel
dwell time. Optical parametric oscillators (OPOs) are a long wavelength (often 1100-1600 nm),
high repetition rate (80 MHz) option with which Kobat et al. demonstrated imaging well past 1
mm in the mouse brain [14]. These are not without their downsides, however. Custom OPOs are
very difficult to build and operate due to their complexity. Commercial oscillators with similar
performance to custom OPOs (Insight, Spectra Physics; Chameleon Discovery, Coherent) can be
purchased that have simple turn-key operation, but they come at a significant price. In this paper
we demonstrate a cost-efficient, simple alternative involving a fiber amplifier and a diamond
Raman laser which produce high power ultrafast outputs at long wavelengths (1060 and 1250 nm
respectively). We demonstrate that the lasers operate at a repetition rate (80 MHz) compatible
with resonant scanning strategies. Furthermore, we demonstrate the capability to use our sources
for neuronal imaging, and vascular imaging to a depth of 1 mm in the mouse neocortex.

2. Materials and methods

2.1. Ultrafast excitation lasers
The excitation lasers used in this work consist of an ytterbium (Yb) fiber amplifier and diamond
Raman laser (Fig. 1(A) and 1(B)). These are more powerful iterations of similar lasers for which
general build instructions are included with a previous publication [15]. The main upgrades
were made to the fiber amplifier and involve the pump and the fiber itself. The pump has been
changed from an 18W to a 30W laser diode (K915FA3RN-30.00W, BWT). The fiber was then
swapped to one that absorbs pump light more efficiently (Liekki Yb1200-25/250DC-PM, nLight).
This ultimately results in a more powerful Yb amplifier, which also increases the diamond
Raman laser output power. For the current amplifier, amplification occurs in the 25 µm core of
an Yb-doped double-clad polarization-maintaining fiber 6 meters in length. Custom FC/APC
connectors were added to each fiber end to prevent damage when operating at high powers
(Coastal Connections, Ventrua, CA). The amplifier is seeded with 75 mW from a commercial
fiber oscillator with an 80 MHz repetition rate (Origami, NKT Photonics). The output from a
laser diode at λ=915 nm serves as the pump, which is propagated backwards relative to the seed
through the 250 µm inner cladding of the fiber. The pump diode power is adjusted to set the
output power of the amplifier. Operation in the parabolic amplification regime ensures that pulses
acquire only a linear chirp, allowing pulse compression with a transmission grating pair upon
output [16–20]. We have achieved an average power as high as 6.5 W following compression,
with pulses centered at λ=1060 nm that are approximately 110 fs in length assuming a sech2
shape (Fig. 2(A)-(C)). Spectra were measured with an NIR spectrometer (AvaSpec-NIR256-1.7,
Avantes) and autocorrelations were recorded using a commercial autocorrelator (pulseCheck,
APE). The amplifier output is directed to either to the diamond Raman laser or to one of
our microscopes at a power established using a half-wave plate and Glan-laser polarizer in
combination.
The diamond Raman laser, which is configured as a ring cavity based on previously published
designs [21,22], is pumped by the fiber amplifier. Mode-matching lenses expand and then focus
the pump to an approximate waist radius of ω0 = 20 µm within the diamond crystal (CVD-Grown,
8mm, AR-coated for λ=1250 nm). The diamond is positioned using a copper stage that is
machined to align the <111> axis of the crystal with the horizontally-polarized pump light from
the fiber amplifier. The pump pulses provide Raman gain for a circulating pulse at the Stokes
wavelength on each round trip. The ring cavity length matches the pulse separation established
by the 80 MHz repetition rate of the fiber amplifier. The cavity is built with high percentage
reflectors at λ=1250 nm (R > 99.5%) apart from one 8% output coupler that passes Stokes light
as the diamond laser output. Cavity length, alignment, and pump pulse width are all balanced
to achieve both a stable and high average power output. We found that using pump pulses with
a slight negative chirp allows such operation, and that the resulting diamond laser pulses are
 Research Article Vol. 13, No. 4 / 1 Apr 2022 / Biomedical Optics Express 1890

Fig. 1. Schematics of the Yb fiber amplifier (A) and diamond Raman laser (B). Also shown
in (C) is a schematic of our multiphoton microscope that employed a resonant scanner. D -
dichroic, ISO - isolator, λ/2 - half-wave plate, GLP - Glan-laser polarizer, OC - output coupler,
SL - scan lens, TL - tube lens, CO - collection optics, EOM - electro-optic modulator.

sufficiently short such that additional dispersion compensation and pulse compression is not
needed for our imaging. We have achieved average powers just over 1.3 W with pulses centered at
λ=1250 nm that are approximately 100 fs in length with a pump power of 6.5 W (Fig. 2(D)-(F)).
Figure 2(G) shows how the diamond Raman laser output changes with pump power. The resulting
slope efficiency is about 31%, comparable with other femtosecond diamond Raman lasers [21,22].
Also shown in Fig. 2(G) is the power of residual pump light that exits the ring cavity after passing
through the diamond crystal for each pumping condition.

2.2. Multiphoton microscopes

Two multiphoton microscopes were used in this work. In both cases, the excitation laser beam
diameter is expanded with a telescope to fill the back aperture of the microscope objective
(XLUMPLFLN 20X, 0.95 NA, Olympus). The first microscope (Fig. 1(C)) will be referred to
as the resonant-galvo microscope [23]. Here, an electro-optic modulator regulates the beam
power sent to the sample (350-80, Conoptics). A resonant scanner (CRS 8kHz, Cambridge)
and a galvanometer scanner (GVS012, Thorlabs) are coupled together with a pair of scan
lenses (SL50-2P2, Thorlabs) and are used to steer the excitation light in a raster pattern. An
additional scan lens and a Plössl tube lens (AC508-400-C, Thorlabs) image these scanners onto
the back aperture of the microscope objective. Fluorescence is epicollected and directed to a
photomultiplier tube (H10770PB-40, Hamamatsu) by a 775 nm cutoff dichroic (FF775-Di01,
Semrock) and is further filtered by a bandpass filter (FF01-609/181-25, Semrock). ScanImage
from Vidrio Technologies controlled the image acquisition [24].
The second microscope will be referred to as the galvo-galvo microscope. For this, we now
use a half-wave plate in combination with a polarizing beam splitter to adjust excitation power. A
pair of galvanometer scanners (6125HB, Cambridge Technology) steer the excitation light. The
previously described scan lens and tube lens combination image the scanners onto the objective
back aperture. Once again, a 775 nm cutoff dichroic guides fluorescence to the detector, which
is now filtered by a 770 nm shortpass filter (FF01-770/SP, Semrock). The detector here is a
photomultiplier tube with superior sensitivity at longer wavelengths (H10770PB-50, Hamamatsu).
 Research Article Vol. 13, No. 4 / 1 Apr 2022 / Biomedical Optics Express 1891

Fig. 2. Characterization of laser system. The top row contains the fiber amplifier
autocorrelation (A), spectrum (B), and power output over 15 minutes (C). The bottom row
similarly contains the autocorrelation (D) and spectrum (E) for the diamond Raman laser.
(F) shows the diamond laser output power over 15 minutes while pumped with the fiber
amplifier at the power shown in (C), using pump pulses possessing a slight negative chirp.
(G) displays the diamond laser output power as a function of pump power from the fiber
amplifier (red) and shows the power of residual pump light that exits the diamond laser cavity
for each pump power (blue-green).

Custom software written with National Instrument’s LabVIEW controlled the image acquisition.
This microscope was also capable of performing laser speckle contrast imaging [25].

2.3. Animal preparation and imaging

Animal procedures were approved by The University of Texas at Austin Institutional Animal
Care and Use Committee. During both surgeries and imaging sessions mice were anesthetized
with isoflurane while body temperature was maintained at 37.5° Celsius. Mice were surgically fit
with cranial windows and allowed to recover for at least one week prior to imaging. To label
neurons with tdTomato, Ai14 mice (#007914, Jackson Laboratory) were intracortically injected
with an AAV5-CaMKII-Cre vector (105558-AAV5, Addgene). Vasculature was labeled through
retro-orbital injections of fluorescent dye diluted in physiological saline. Excitation power was
minimized when imaging superficial regions and increased with depth. Power incident on the
brain surface was maintained under 250 mW to avoid tissue damage [26]. The only instance
where excitation powers approached this threshold was during the deep imaging experiments
covered in section 3.2.
 Research Article Vol. 13, No. 4 / 1 Apr 2022 / Biomedical Optics Express 1892

3. Experimental results
3.1. In vivo imaging with a resonant scanner
Vasculature across a large (>1 mm2 ) lateral field of view was imaged with the resonant-galvo
microscope using the fiber amplifier for excitation. The grating compressor was set to minimize
the pulse width at the focal plane prior to imaging. 100 µL of 70 kDa dextran-conjugated Texas
Red diluted in saline at a 5% w/v ratio was added to the blood plasma prior to the imaging session
to serve as the label. 20 frames were averaged at each depth to produce images, which each
originally had a 700 by 700 µm field of view. Each stack extended to a depth of 660 µm with a 3
µm step size in between slices. 4 stacks were acquired in adjacent regions and stitched together
afterwards using ImageJ [27]. The final stitched mosaic image had a 1160 by 1160 µm lateral
field of view. Figure 3(A) shows a maximum intensity projection through the entire mosaic stack.
Figure 3(B) likewise contains a side projection through the stack. A laser speckle contrast image
of blood flow on the cortical surface of the imaged mouse is shown in Fig. 3(C), where a box was
added to indicate the region over which MPM was performed. Matching vascular features are
easily observed in both the MPM and the laser speckle contrast images.

Fig. 3. Imaging of Texas Red-labeled vasculature using the Yb fiber amplifier along with
a resonant scanner. The lateral field of view is 1160 by 1160 µm. (A) is the overhead
maximum intensity projection through the stack, (B) contains the side projection, and (C) is
a laser speckle contrast image that depicts the surface vasculature of the imaged brain region.
Scale bars are 200 µm. Post objective power ranged from 10-130 mW.

When imaging with a resonant scan mirror, individual red blood cells can be captured in single
frames of vascular images due to the considerably reduced acquisition time [28,29]. We show
this in Fig. 4, which contains both a frame-averaged image of vascular structure (Fig. 4(A)) along
with a single frame from those averaged (Fig. 4(B)). Locations of red blood cells in the plasma
are clearly seen as shadows in Fig. 4(B). A single vessel in the network was imaged repeatedly
to thoroughly evaluate red blood cell motion (Fig. 4(C), Visualization 1). The resonant-galvo
microscope can acquire images with 512 by 512 pixels at a ∼28 Hz frame rate. By decreasing the
image to 512 by 30 pixels, however, this is increased to ∼184 Hz. With the improved sampling
speed, red blood cell velocity was determined from a kymograph (Fig. 4(D)) where the intensity
along the line drawn in Fig. 4(C) was plotted for each image of the single vessel. The slope of
 Research Article Vol. 13, No. 4 / 1 Apr 2022 / Biomedical Optics Express 1893

the resulting streaks indicates a red blood cell velocity of 0.67 ± 0.15 mm/sec, which is within
the standard range for capillaries [30,31].

Fig. 4. Imaging red blood cells with a resonant scanner. Plasma is labeled with Texas Red,
and images were acquired at a ∼400 µm depth. (A) is a frame averaged image of vascular
structure, whereas (B) is a single frame demonstrating the ability to image red blood cells.
Scale bar is 20 µm. (C) is a single-frame image of the vessel in the center of (A) and (B),
which was imaged repeatedly at a 184 Hz rate. (D) is a kymograph created by plotting the
intensity across the line drawn through the center of the vessel in (C) for each acquired
image. The horizontal scale bar is 10 µm, the vertical scale bar is 100 msec. Using (D), a
0.67 ± 0.15 mm/sec velocity is observed.

3.2. Deep imaging with the diamond Raman laser

The diamond Raman laser’s spectral properties make it suitable for deep imaging when paired
with red-shifted bright dyes. In Fig. 5 vasculature is imaged to just beyond 1 mm in the brain by
exciting Alexa Fluor 647 (80 µL injection at a concentration of 5% w/v) and Alexa Fluor 680
(100 µL injection at a concentration of 5% w/v). The Alexa Fluor 647 image stack has a 5 µm
step size between slices, whereas the Alexa Fluor 680 stack has a 3 µm step size. Both stacks are
composed of images with a 400 by 400 µm field of view. Frame averaging in superficial regions
was limited to 3 frames per slice, but this was increased to as high as 10 frames in the deeper
regions of the stacks. Excitation power was minimized at the surface and increased with depth,
but the average power incident on the brain surface was not allowed to reach 250 mW. Towards
the left of Fig. 5 is the side projection and various overhead maximum intensity projections
through the images of vasculature labeled with Alexa Fluor 680. Towards the right of Fig. 5 is
the same content for the vasculature labeled with Alexa Fluor 647. Note the large penetrating
vessel that appears at a depth of about 500 µm for the Alexa Fluor 680 images. If this did not
appear, we would expect to resolve the vasculature towards the left side of the projection from
900 to 1056 µm just as well as on the right.

3.3. Neuronal and vascular imaging

Neuronal and vascular images from the same cortex location in a mouse were acquired separately
during a single session using the galvo-galvo microscope (Fig. 6). Pyramidal neurons expressing
tdTomato and vasculature labeled with an 80 µL retro-orbital injection of 10 kDa dextran-
conjugated Alexa Fluor 647 diluted in saline at a 2.5% w/v ratio were imaged. Signal from
 Research Article Vol. 13, No. 4 / 1 Apr 2022 / Biomedical Optics Express 1894

Fig. 5. Deep vascular imaging of Alexa Fluor 680 (left) and Alexa Fluor 647 (right) with the
diamond Raman laser. Both side projections and overhead maximum intensity projections
are displayed. The lateral field of view for the images is 400 by 400 µm. Scale bars are 100
µm. Post objective power ranged from <10 mW at the brain surface, to 120 mW at 800 µm,
to 220 mW at a 1 mm depth.

one fluorophore versus the other could be discerned based on the excitation source since Alexa
Fluor 647 is efficiently excited by only the diamond Raman laser at λ=1250 nm and tdTomato
is efficiently excited by only the fiber amplifier at λ=1060 nm [32,33]. First, vascular images
were acquired using the diamond Raman laser. Images were recorded to a depth of 765 µm using
a 3 µm step size between slices, and 3 frames were averaged to create each image in the stack.
Imaging of the tdTomato-labeled neurons using the fiber amplifier followed the acquisition of the
vascular stack. Images were acquired to a depth of 765 µm with a 3 µm step size, and 5 frames
were averaged together at each depth. The lateral field of view for the vascular imaging was ∼400
by 400 µm, whereas the neuronal imaging field of view was 235 by 235 µm. The vascular images
shown in Fig. 6 have been cropped to match the neuronal field of view. Since vascular structure
is seen in many neuronal images as regions of reduced signal, the two stacks were merged after
acquisition by using a MATLAB program to overlap common vascular features. Figure 6(A)
shows maximum intensity projections from the surface to a depth of 450 µm for the neuronal
stack, the vascular stack, and the merged stack. Figure 6(B) contains side projections through
the entire stacks for each structure. Axons stemming from the neuronal cell bodies are clearly
seen in the side projections. Figure 6(C) contains some projections through various depths of the
merged stack. Dendrites can be seen stemming from neurons towards superficial regions, and
neuronal cell bodies can be seen throughout the entire depth. A complete fly-through of all the
stacks can be found in Visualization 2.
 Research Article Vol. 13, No. 4 / 1 Apr 2022 / Biomedical Optics Express 1895

Fig. 6. Neuronal and vascular images. (A) Maximum intensity projections spanning
from the surface to a depth of 450 µm for neurons (left, tdTomato, λ=1060 nm excitation),
vasculature (right, Alexa Fluor 647, λ=1250 nm excitation), and the merged stack (middle).
(B) Side projections for the same stacks from (A). (C) Maximum intensity projections at
different depths within the merged stack. Scale bars are 50 µm.

4. Discussion
The fiber amplifier and diamond Raman laser presented here offer many unique opportunities.
Most notably, they produce high output powers with a relatively simple design that is straightfor-
ward to implement. The overall design is similar to our previous work [15] and the modifications
made did not significantly add to the system cost, which was still ∼$50,000 for all components
including the seed laser. The new design has increased output power for the Yb amplifier and
the diamond laser respectively to 6.5 W at λ=1060 nm and 1.3 W at λ=1250 nm. These power
and wavelength combinations allow deep imaging while operating at a repetition rate (80 MHz)
that enables fast image acquisition using resonant scanners. When adopting a fast-scanning
strategy it is important to ensure the excitation laser has a repetition rate that delivers multiple
pulses within the shortened pixel dwell time. In Fig. 3 we demonstrate that our lasers can do this
when paired with our 8 kHz resonant scanner. Note that 20 frames were averaged together in
the images acquired for this figure, while we limited averaging to 3 frames for similar images
acquired using galvanometer-driven scanning. Despite this it is still more than 3 times faster
to record stacks with the resonant scanner. Decreasing imaging time reduces the possibility of
 Research Article Vol. 13, No. 4 / 1 Apr 2022 / Biomedical Optics Express 1896

damaging the brain through prolonged exposure to excitation light or harming mice through
excessive anesthetic inhalation. It also reduces the magnitude of signal reduction throughout a
vascular imaging session due to fluorophore clearance from the bloodstream. Addressing these
issues may prove beneficial for chronic studies involving a large field of view that are aided
by resonant scanning. Note that while we did not test the diamond Raman laser with resonant
scanning, we do not foresee any issues with this.
While we did not use the diamond Raman laser with resonant scanning, we did make use of
this laser to image relatively deep brain regions. Both excitation sources in our system produce
pulses compressible to relatively short pulse widths (∼100 fs). This, along with their high average
powers (both >1W) and relatively long wavelengths, enables deep imaging which we demonstrate
in Fig. 5. Here, dispersion compensation was not required to image the entire depth of the
cerebral cortex (∼1 mm) though two-photon excitation (2PE) using the diamond Raman laser.
That said, adding compensation to minimize pulse width in the imaging plane could potentially
reduce the average excitation power required to resolve deep structure. Compensation could
also improve the ability to excite fluorophores through three-photon excitation (3PE). 3PE of
fluorophores at λ=1250 nm has been demonstrated, but this requires a high peak power as 3PE
is a less efficient process compared to 2PE [34,35]. 3PE can improve imaging in deep tissue
locations due to decreased background fluorescence, but lower repetition rate sources are often
used for this technique to maximize pulse energy and reduce average power [36,37].
When compared to other common high repetition rate excitation sources such as Ti:S lasers,
both the fiber amplifier and diamond Raman laser have longer wavelength outputs. To take
advantage of the decreased scattering offered by the long wavelengths of our lasers, fluorophores
that are efficiently excited by each source must be identified. The λ=1250 nm output of the
diamond Raman laser pairs well with both Alexa Fluor 647 and Alexa Fluor 680 as shown in
Fig. 5. The λ=1060 nm output of the fiber amplifier pairs well with Texas Red for vascular
imaging (Fig. 3, Fig. 4), and tdTomato which we used to label neurons (Fig. 6). Also important,
the spectral difference between the two lasers is sufficient such that the fluorophores excited by
one source are not easily excited by the other. This allows us to image multiple structures in
a single sample if they are labeled with different fluorophores. As an example, we separately
imaged vasculature labeled with Alexa Fluor 647 and then neurons labeled with tdTomato during
the same session to create the merged images shown in Fig. 6. In addition to the labels previously
mentioned, there is currently a push to develop red-shifted fluorophores that are excitable by
far-red wavelengths like what our lasers provide [38–40]. The utility of the laser system will
expand as new labels are introduced. Our system can also excite fluorophores at intermediate
effective wavelengths if used to initiate non-degenerate excitation (NDE) [15,41–43].
One limitation of the laser pair presented here is their fixed wavelengths. Alternatives such as
OPOs and commercial laser systems (Insight, Spectra Physics; Chameleon Discovery, Coherent)
have tunable outputs (∼1100-1600 nm for OPOs, 700-1300 nm for commercial systems) and
similarly high repetition rates. In practice shorter wavelengths are less useful for deeper in vivo
imaging due to increased scattering, and wavelengths between 1400 and 1600 nm are rarely
used for MPM due to increased water absorption in tissue [12]. Tuning an OPO to a specific
wavelength can maximize excitation efficiency for a given fluorophore, but this ability adds
complexity to the instrumentation. When using a custom OPO, exact phase matching conditions
must be set to produce a desired wavelength. The fiber amplifier and diamond laser system is
much simpler in that it is for the most part turn-key. Building the laser system is not significantly
more complex than building a custom multiphoton scanning microscope. After initial alignment
only the diamond laser’s cavity length needs to be routinely adjusted. Therefore, the system
presented in this paper is a cost-efficient, long wavelength multiphoton excitation source with
high output powers that can be constructed by labs with some optics expertise.
 Research Article Vol. 13, No. 4 / 1 Apr 2022 / Biomedical Optics Express 1897

5. Conclusion
MPM is rapidly improving as new technologies and laser systems for the imaging strategy are
developed. We have introduced a set of custom excitation lasers for MPM that are valuable for
imaging neural structures. The lasers are an Yb fiber amplifier and a diamond Raman laser with
high power outputs (6.5 W, 1.3 W) at wavelengths (1060 nm, 1250 nm) excellent for exciting many
red-shifted neural labels. We have shown that neurons and vasculature in the mouse cerebral
cortex can be reliably imaged with the system. Vascular structure was imaged to depths greater
than 1 mm. Additionally, enabled by their high repetition rate (80 MHz), we have shown that
our lasers are compatible with a resonant scanning technique that allows fast image acquisition.
The laser system we developed is a simple and cost-efficient alternative to many other common
custom and commercial excitation sources.
Funding. National Institutes of Health (EB011556, NS108484, T32EB007507, T32LM012414); UT Austin Portugal
Program.
Disclosures. The authors declare no conflicts of interest.
Data availability. Data underlying the results presented in this paper are not publicly available at this time but may
be obtained from the authors upon reasonable request.

References
1. K. Svoboda and R. Yasuda, “Principles of two-photon excitation microscopy and its applications to neuroscience,”
Neuron 50(6), 823–839 (2006).
2. A. Y. Shih, J. D. Driscoll, P. J. Drew, N. Nishimura, C. B. Schaffer, and D. Kleinfeld, “Two-photon microscopy
as a tool to study blood flow and neurovascular coupling in the rodent brain,” J. Cereb. Blood Flow Metab. 32(7),
1277–1309 (2012).
3. C. Stosiek, O. Garaschuk, K. Holthoff, and A. Konnerth, “In vivo two-photon calcium imaging of neuronal networks,”
Proc. Natl. Acad. Sci. 100(12), 7319–7324 (2003).
4. C. Xu and F. Wise, “Recent advances in fibre lasers for nonlinear microscopy,” Nature photon 7(11), 875–882 (2013).
5. F. Helmchen and W. Denk, “Deep tissue two-photon microscopy,” Nat. Photonics 2(12), 932–940 (2005).
6. E. Beaurepaire, M. Oheim, and J. Mertz, “Ultra-deep two-photon fluorescence excitation in turbid media,” Opt.
Commun. 188(1-4), 25–29 (2001).
7. P. Theer, M. T. Hasan, and W. Denk, “Two-photon imaging to a depth of 1000 µm in living brains by use of a
Ti:Al2 O3 regenerative amplifier,” Opt. Lett. 28(12), 1022–1024 (2003).
8. D. R. Miller, A. M. Hassan, J. W. Jarrett, F. A. Medina, E. P. Perillo, K. Hagan, S. M. Kazmi, T. A. Clark, C.
T. Sullender, T. A. Jones, B. V. Zemelman, and A. K. Dunn, “In vivo multiphoton imaging of a diverse array of
fluorophores to investigate deep neurovascular structure,” Biomed. Opt. Express 8(7), 3470–3481 (2017).
9. D. G. Ouzounov, T. Wang, M. Wang, D. D. Feng, N. G. Horton, J. C. Cruz-Hernández, Y. T. Cheng, J. Reimer, A. S.
Tolias, N. Nishimura, and C. Xu, “In vivo three-photon imaging of activity of GCaMP6-labeled neurons deep in
intact mouse brain,” Nat. Methods 14(4), 388–390 (2017).
10. M. Yildirim, H. Sugihara, P. T. C. So, and M. Sur, “Functional imaging of visual cortical layers and subplate in awake
mice with optimized three-photon microscopy,” Nat. Commun. 10(1), 177 (2019).
11. D. Kobat, M. E. Durst, N. Nishimura, A. W. Wong, C. B. Schaffer, and C. Xu, “Deep tissue multiphoton microscopy
using longer wavelength excitation,” Opt. Express 17(16), 13354–13364 (2009).
12. D. R. Miller, J. W. Jarrett, A. M. Hassan, and A. K. Dunn, “Deep tissue imaging with multiphoton fluorescence
microscopy,” Curr. Opin. Biomed. Eng. 4, 32–39 (2017).
13. N. Ji, J. Freeman, and S. L. Smith, “Technologies for imaging neural activity in large volumes,” Nat. Neurosci. 19(9),
1154–1164 (2016).
14. D. Kobat, N. G. Horton, and C. Xu, “In vivo two-photon microscopy to 1.6-mm depth in mouse cortex,” J. Biomed.
Opt. 16(10), 106014 (2011).
15. E. P. Perillo, J. W. Jarrett, Y. Liu, A. Hassan, D. C. Fernée, J. R. Goldak, A. Bonteanu, D. J. Spence, H. Yeh, and A.
K. Dunn, “Two-color multiphoton in vivo imaging with a femtosecond diamond Raman laser,” Light: Sci. Appl.
6(11), e17095 (2017).
16. V. I. Kruglov, A. C. Peacock, J. D. Harvey, and J. M. Dudley, “Self-similar propagation of parabolic pulses in
normal-dispersion fiber amplifiers,” J. Opt. Soc. Am. B 19(3), 461–469 (2002).
17. J. M. Dudley, C. Finot, D. J. Richardson, and G. Millot, “Self-similarity in ultrafast nonlinear optics,” Nat. Phys. 3(9),
597–603 (2007).
18. J. Limpert, T. Schreiber, T. Clausnitzer, K. Zöllner, H-J. Fuchs, E-B. Kley, H. Zellmer, and A. Tünnermann,
“High-power femtosecond Yb-doped fiber amplifier,” Opt. Express 10(14), 628–638 (2002).
 Research Article Vol. 13, No. 4 / 1 Apr 2022 / Biomedical Optics Express 1898

19. D. N. Papadopoulos, Y. Zaouter, M. Hanna, F. Druon, E. Mottay, E. Cormier, and P. Georges, “Generation of 63 fs
4.1 MW peak power pulses from a parabolic fiber amplifier operated beyond the gain bandwidth limit,” Opt. Lett.
32(17), 2520–2522 (2007).
20. Y. Deng, C. Chien, B. G. Fidric, and J. D. Kafka, “Generation of sub-50 fs pulses from a high-power Yb-doped fiber
amplifier,” Opt. Lett. 34(22), 3469–3471 (2009).
21. M. Murtagh, J. Lin, R. P. Mildren, and D. J. Spence, “Ti: sapphire-pumped diamond Raman laser with sub-100-fs
pulse duration,” Opt. Lett. 39(10), 2975–2978 (2014).
22. M. Murtagh, J. Lin, R. P. Mildren, G. McConnell, and D. J. Spence, “Efficient diamond Raman laser generating 65 fs
pulses,” Opt. Express 23(12), 15504–15513 (2015).
23. A. Zhou, S. A. Engelmann, S. A. Mihelic, A. Tomar, A. M. Hassan, and A. K. Dunn, “An evaluation of resonant
scanning as a high-speed imaging technique for two-photon imaging of cortical vasculature,” BioRXiv 13(3),
1374–1385 (2021).
24. T. A. Pologrunto, B. L. Sabatini, and K. Svoboda, “ScanImage: flexible software for operating laser scanning
microscopes,” BioMed Eng OnLine 2(1), 1–9 (2003).
25. C. J. Schrandt, S. M. Kazmi, T. A. Jones, and A. K. Dunn, “Chronic monitoring of vascular progression after ischemic
stroke using multiexposure speckle imaging and two-photon fluorescence microscopy,” J. Cereb. Blood Flow Metab.
35(6), 933–942 (2015).
26. K. Podgorski and G. Ranganathan, “Brain heating induced by near-infrared lasers during multiphoton microscopy,” J.
Neurophysiol. 116(3), 1012–1023 (2016).
27. S. Preibisch, S. Saalfeld, and P. Tomancak, “Globally optimal stitching of tiled 3D microscopic image acquisitions,”
Bioinformatics 25(11), 1463–1465 (2009).
28. J. L. Fan, J. A. Rivera, W. Sun, J. Peterson, H. Haeberle, S. Rubin, and N. Ji, “High-speed volumetric two-photon
fluorescence imaging of neurovascular dynamics,” Nat. Commun. 11(1), 1–12 (2020).
29. B. Li, C. Wu, M. Wang, K. Charan, and C. Xu, “An adaptive excitation source for high-speed multiphoton microscopy,”
Nat. Methods 17(2), 163–166 (2020).
30. P. J. Drew, A. Y. Shih, and D. Kleinfeld, “Fluctuating and sensory-induced vasodynamics in rodent cortex extend
arteriole capacity,” Proc. Natl. Acad. Sci. 108(20), 8473–8478 (2011).
31. J. Y. Huang, L. T. Li, H. Wang, S. S. Liu, Y. M. Lu, M. H. Liao, R. R. Tao, L. J. Hong, K. Fukunaga, Z. Chen, C. S.
Wilcox, E. Y. Lai, and F. Han, “In vivo two-photon fluorescence microscopy reveals disturbed cerebral capillary blood
flow and increased susceptibility to ischemic insults in diabetic mice,” CNS Neurosci. Ther. 20(9), 816–822 (2014).
32. A. Rakhymzhan, R. Leben, H. Zimmermann, R. Günther, P. Mex, D. Reismann, C. Ulbricht, A. Acs, A. U. Brandt,
R. L. Lindquist, T. H. Winkler, A. E. Hauser, and R. A. Niesner, “Synergistic strategy for multicolor two-photon
microscopy: application to the analysis of germinal center reactions in vivo,” Sci. Rep. 7(1), 7101 (2017).
33. M. Drobizhev, S. Tillo, N. S. Makarov, T. E. Hughes, and A. Rebane, “Absolute two-photon absorption spectra and
two-photon brightness of orange and red fluorescent proteins,” J. Phys. Chem. B 113(4), 855–859 (2009).
34. A. A. Lanin, M. S. Pochechuev, A. S. Chebotarev, I. V. Kelmanson, D. S. Bilan, D. A. Kotova, V. S. Tarabykin, A. A.
Ivanov, A. B. Fedotov, V. V. Belousov, and A. M. Zheltikov, “Cell-specific three-photon-fluorescence brain imaging:
neurons, astrocytes, and gliovascular interfaces,” Opt. Lett. 45(4), 836–839 (2020).
35. D. G. Ouzounov, T. Wang, C. Wu, and C. Xu, “GCaMP6 ∆F/F dependence on the excitation wavelength in 3-photon
and 2-photon microscopy of mouse brain activity,” Biomed. Opt. Express 10(7), 3343–3352 (2019).
36. T. Wang and C. Xu, “Three-photon neuronal imaging in deep mouse brain,” Optica 7(8), 947–960 (2020).
37. N. G. Horton, K. Wang, D. Kobat, C. G. Clark, F. W. Wise, C. B. Schaffer, and C. Xu, “In vivo three-photon
microscopy of subcortical structures within an intact mouse brain,” Nat. Photonics 7(3), 205–209 (2013).
38. M. Inoue, A. Takeuchi, S. I. Horigane, M. Ohkura, K. Gengyo-Ando, H. Fujii, S. Kamijo, S. Takemoto-Kimura, M.
Kano, J. Nakai, and K. Kitamura, “Rational design of a high-affinity, fast, red calcium indicator R-CaMP2,” Nat.
Methods 12(1), 64–70 (2015).
39. C. Tischbirek, A. Birkner, H. Jia, B. Sakmann, and A. Konnerth, “Deep two-photon brain imaging with a red-shifted
fluorometric Ca2 + indicator,” Proc. Natl. Acad. Sci. 112(36), 11377–11382 (2015).
40. H. Dana, B. Mohar, Y. Sun, S. Narayan, A. Gordus, J. P. Hasseman, G. Tsegaye, G. T. Holt, A. Hu, D. Walpita, and R.
Patel, “Sensitive red protein calcium indicators for imaging neural activity,” eLife 5, e12727 (2016).
41. S. Sadegh, M. Yang, C. G. L. Ferri, M. Thunemann, P. A. Saisan, A. Devor, and Y. Fainman, “Measurement of
the relative non-degenerate two-photon absorption cross-section for fluorescence microscopy,” Opt. Express 27(6),
8335–8347 (2019).
42. S. Sadegh, M. Yang, C. G. L. Ferri, M. Thunemann, P. A. Saisan, Z. Wei, E. A. Rodriguez, S. R. Adams, K. Kiliç, D.
A. Boas, S. Sakadžić, A. Devor, and Y. Fainman, “Efficient non-degenerate two-photon excitation for fluorescence
microscopy,” Opt. Express 27(20), 28022–28035 (2019).
43. P. Mahou, M. Zimmerley, K. Loulier, K. S. Matho, G. Labroille, X. Morin, W. Supatto, J. Livet, D. Débarre, and E.
Beaurepaire, “Multicolor two-photon tissue imaging by wavelength mixing,” Nat. Methods 9(8), 815–818 (2012).