Quantum cascade laser-based CO2 isotope sensing: Devices, systems, and implications

Matthew Escarra*a,b, Loan Leb, Stephen Soc,

David Thomazyc, Michael Oppenheimerd, and Claire Gmachlb

aDept. of Applied Physics, California Institute of Technology, Pasadena, CA, USA 91125; bDept. of Electrical Engineering, Princeton University, Princeton, NJ USA 08544;

cSentinel Photonics, Plainsboro, NJ USA 08536; dWilson School of Public and International Affairs, Princeton University, Princeton, NJ USA 08544

BIOGRAPHY Matthew Escarra is a Postdoctoral Scholar in Applied Physics and Materials Science at the California Institute of Technology working with Prof. Harry Atwater. At Caltech, Matthew is developing spectrum splitting photonic architectures and solar cells for high efficiency solar energy conversion.

Matthew received a PhD in Electrical Engineering at Princeton University, in the group of Prof. Claire Gmachl, where he focused on high performance quantum cascade lasers and mid-infrared metamaterials. Prior to Princeton, Matthew earned a BS in Electrical Engineering at Rice University. Matthew has also worked at Daylight Solutions and Sentinel Photonics, two start-ups in the mid-infrared optoelectronics community.

TECHNICAL ABSTRACT CO2 fingerprinting via isotope ratio measurements offers the potential to address several important technological and societal issues. Tracking the relative concentrations of 12CO2, 13CO2, and 18OCO (the CO2 fingerprint) across different environments can improve our understanding of the carbon cycle. A sensor network can be used to measure spatial and temporal fluxes of CO2 isotopes for monitoring emissions and verifying abatement from particular sites. In enhanced oil recovery or carbon capture and storage (CCS), a network of sensors can monitor the area surrounding underground reservoirs for leaks, by searching for the fingerprint corresponding with the injected CO2.

In this talk, I will discuss the design and development of distributed feedback QC lasers for trace gas detection of strong lines of the less abundant 13CO2 and 18OCO and a weaker line of the abundant 12CO2. I will then highlight the development of a QC-laser based sensor for CO2 isotope detection at Sentinel Photonics. Finally, I will discuss the potential economic impact of this type of sensing technology, quantified via an integrated climate-economy model.

Unlike conventionally used isotope ratio mass spectrometry [1], a laser-based sensor allows real-time, in situ, continuous monitoring of CO2 isotopes in a compact and autonomous format, enabling large-scale sensor networks that provide spatially and temporally resolved CO2 isotope flux measurements. Quantum cascade (QC) lasers are the ideal

0.0 0.2 0.4 0.6 0.8 1.00

2

4

6

8

10

12

14

16

Normal Atmosphere N

2 Atmosphere

CO2 Rich Atmosphere

Current (A)

Vol

tage

(V)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Pow

er (W

)

Figure 1: a) Absorption cross-section spectra for the 12CO2, 13CO2, and 18OCO isotopes of CO2 from the HITRAN2008 database [3]. A single mode, CW QC laser spectrum is overlaid to show proximity to strong absorption lines of all three isotopologues. b) CW LIV data is shown for the same device tested in three different atmospheric conditions.

b) a)

*escarra@caltech.edu

choice for such measurements, as their mid-infrared emission overlaps well with strong CO2 vibrational absorption lines. I will discuss two QC laser design approaches for use in this application. The first QC lasers used in this study feature a low voltage defect and short injector in their quantum design [2]. To achieve single-mode operation, distributed feedback (DFB) gratings were etched 500 nm into the top of these devices with a grating period of 680 nm. An example continuous wave (CW) spectrum for one of these lasers is plotted in Fig. 1a along with HITRAN absorption cross-section spectra for 12CO2, 13CO2, and 18OCO [3]. One can see that this representative device has a single lasing mode near absorption peaks for all three isotopes of CO2. To confirm the CO2 sensing ability of these devices, LIV measurements were compared in different atmospheres, as shown in Fig. 1b. Measurements in a typical room atmosphere result in strong dips in the light-current curve whenever the laser mode passes over a CO2 absorption peak. When the laser is operated in an environment purged with dry nitrogen, these dips disappear. When the laser is run in a CO2 saturated atmosphere created by the sublimation of dry ice, virtually no light makes it to the detector. A second QC laser, featuring strong injector-upper laser level coupling and very highly strained active region layers, was developed with emission at 4.34 μm, providing similar spectral overlap with all three of these isotopes of CO2.

Through collaboration with Sentinel Photonics, a compact and efficient CO2 isotope trace gas sensor was developed featuring analogous QC lasers in a direct absorption measurement. This sensor is designed to be the size of a shoe box, light in weight (< 15 pounds), and low in power consumption (< 20 W). A QC laser is beam-split and coupled to two 3.5 m path length multipass cells (a sample cell and a reference cell). The laser light passes through the sample and is then measured by a thermoelectrically cooled MCT detector. The gas sample’s temperature and pressure were well controlled, in order to minimize related noise. Fig. 2a shows a block diagram of this CO2 isotope sensor. The resulting absorption spectrum from this sensor shows strong overlap with HITRAN simulations. Without optimization, the standard deviation of the residual between the sensor performance and the simulation is 0.5%. The stronger line corresponds to 12CO2, and the weaker line corresponds to 13CO2.

To quantitatively assess the economic importance of such a sensor for CO2 leakage detection from CCS reservoirs, we modified the Dynamic Integrated Climate-Economy (DICE) model to include CCS. A study was performed to analyze the sensitivity of model outcomes to reservoir leakage (or equivalently, to the reservoir retention half-life). The model assumes that cost-effective approaches will be adopted. The present day economic value of CCS was established versus reservoir half-life, showing a significant return on investment (~2 trillion U.S.$, or ~4% of gross world product (GWP)) when the average half-life is greater than 250 years, with a sharp drop in the value of CCS technology for half-life values below 250 years (Fig. 2c). This model shows that, in order to reduce uncertainty associated with present-day CCS investments, CCS reservoir leakage rates must be quantified. QC laser-based sensors have the potential to quantify these leakage rates and to assure the safety of these reservoirs for years to come.

This work is supported in part by MIRTHE (NSF-ERC, # EEC-0540832), DARPA-EMIL, the NSF SBIR Program (# 1113289), and the PEI-STEP program at Princeton.

[1] B. Vaughn, C. Evans, J White, C. Still, K. Masarie, and J. Turnbull, “Global network measurements of atmospheric trace gas isotopes” In: Isoscapes, Understanding movement, pattern, and process on Earth through isotope mapping, Springer, Amsterdam, 3–31 (2009). [2] M. Escarra, A. Hoffman, K. Franz, S. Howard, R. Cendejas, X. Wang, J. Fan, and C. Gmachl, “Quantum cascade lasers with voltage defect of less than one longitudinal optical phonon energy,” Appl. Phys. Lett. 94, 251114 (2009). [3] L. Rothman, et al., “The HITRAN 2008 molecular spectroscopic database,” J. Quant. Spectrosc. Radiat. Transfer 110, 533-572 (2009).

Keywords: CO2 isotope, quantum cascade laser, trace gas sensing, carbon capture and storage

Figure 2: a) Block diagram of the CO2 isotope sensor developed at Sentinel Photonics. b) Comparison between HITRAN simulation (blue) and experimental data (red) from this isotope sensor. The lower plot shows the fractional difference between these curves. c) Model results showing the present value of CCS technology with respect to reservoir half-life (CO2 leakage rate) relative to three baseline scenarios.

2302 2302.1 2302.2 2302.3 2302.4 2302.5 2302.6 2302.7 2302.8 2302.90

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Abs

orpt

ion

2302 2302.1 2302.2 2302.3 2302.4 2302.5 2302.6 2302.7 2302.8 2302.9-0.05

0

0.05

wavenumber [cm-1]

Res

idua

l

hitranexp data

13CO

12CO

c) b) a)