Skip to main content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Anal Sci Adv. 2024 Apr; 5(3-4): 2300053.
Published online 2024 Jan 25. doi: 10.1002/ansa.202300053
PMCID: PMC11142394
PMID: 38827022

Two common pitfalls in the analysis of water‐stable isotopologues with cryogenic vacuum extraction and cavity ring‐down spectroscopy

Simon Haberstroh,corresponding author 1 Angelika Kübert, 1 , 2 and Christiane Werner 1
Author information Article notes Copyright and License information PMC Disclaimer

Associated Data

Data Availability Statement

Abstract

Water stable isotopologue analysis is widely used to disentangle ecohydrological processes. Yet, there are increasing reports of measurement uncertainties for established and emerging methods, such as cryogenic vacuum extraction (CVE) or cavity ring‐down spectroscopy (CRDS). With this study, we investigate two pitfalls, that potentially contribute to uncertainties in water‐stable isotopologue research. To investigate fractionation sources in CVE, we extracted pure water of known isotopic composition with cotton, glass wool or without cover and compared the isotopologue results with non‐extracted reference samples. To characterise the dependency of δ2H and δ18O on the water mixing ratio in CRDS, which is of high importance for in‐situ applications with large natural variations in mixing ratios, we chose samples with a large range of isotopic compositions and determined δ2H and δ18O for different water mixing ratios with two CRDS analysers (Picarro, Inc.). Cotton wool had a strong fractionation effect on δ2H values, which increased with more 2H‐enriched samples. δ2H and δ18O values showed a strong dependency on the water mixing ratio analysed with CRDS with differences of up to 34.5‰ (δ2H) and 3.9‰ (δ18O) for the same sample at different mixing ratios. CVE and CRDS, now routinely applied in water stable isotopologue research, come with pitfalls, namely fractionation effects of cover materials and water mixing ratio dependencies of δ2H and δ18O, which can lead to erroneous isotopologue results and thus, invalid conclusions about (ecohydrological) processes. These practical issues identified here should be reported and addressed adequately in water‐stable isotopologue research.

Keywords: cotton, CRDS, Isotopic fractionation, δ18O, δ2H

Abbreviations:

CRDS
Cavity ring‐down spectroscopy
CVE
Cryogenic vacuum extraction

1. INTRODUCTION

Water‐stable isotopologues are a powerful tool to trace a variety of ecohydrological processes, such as plant water uptake or soil water infiltration. 1 , 2 , 3 , 4 To extract water from plant and soil material, cryogenic vacuum extraction (CVE) 5 , 6 is still widely used, 7 although new methods have been developed. 2 , 8 , 9 , 10 , 11 , 12 In addition, recent studies have questioned the validity of CVE due to extraction biases in δ2H values 13 , 14 and problematic comparability between laboratories. 15 Potential reasons for extraction biases include different extraction times, 7 , 16 , 17 , 18 extraction temperature, 19 , 20 sample preparation protocols or leaky vacuum systems. 15 It has also been demonstrated that soil type, composition and water content can have a significant impact on water stable isotopologue composition when CVE is applied. 19 , 21 , 22 , 23 , 24

While the literature is abundant on these issues, a rarely reported problem in CVE is the material, which is used to cover the plant or soil sample to prevent particles from being drawn into the vacuum pump or the extracted water sample. This material is present during the entire extraction process, experiences the same (temperature) conditions and is passed by the extracted water vapour on the way to the cold trap. Therefore, inert glass wool is generally suggested and used. 16 , 22 , 25 However, many studies do not report the used covering materials in CVE. This raises the question of whether glass wool or other materials are used, which could have severe consequences for isotopologue analysis if they are of an organic nature. Indeed, some studies report the use of other (organic) cover materials, such as fleece 7 or cotton wool, 17 although, for example, cotton contains hydrogen and oxygen itself and is considered hygroscopic. 22 , 26 , 27 In addition, cellulose is the major constituent of cotton wool. 28 , 29 Wassenaar and Hobson 30 experimentally determined ∼23% of hydrogen in cotton to be exchangeable at 130°C, which is close to the temperature used in CVE.

Thus, it is likely that there are contamination or fractionation effects of such organic (cover) materials on isotopic composition during CVE. Similar fractionation/contamination effects have been described in other studies, which were often related to the presence of organic materials with exchangeable hydrogen, such as plant material itself. 13 , 30 , 31 , 32

After CVE, extracted water samples are often analysed with laser spectrometers, such as cavity ring‐down spectroscopy (CRDS) instruments. While these instruments offer a fast and inexpensive measurement of water‐stable isotopologues, they come with other issues compared to isotope‐ratio mass spectroscopy, such as organic contaminations 33 , 34 , 35 , 36 or water mixing ratio dependencies. 37 , 38 For example, Picarro Inc., one of the main manufacturers of CRDS analysers, states that one of the current models (L2130‐i) has a recommended measurement range of 1000–50,000 ppm. 39 Yet, it has been demonstrated that the analysed water mixing ratio in CRDS has a strong impact on obtained δ2H and δ18O values, 37 , 38 , 40 which has high relevance for studies, where the isotopic composition is measured directly (and is difficult to control) in the water vapour, such as soil evaporation, soil water vapour, atmospheric water vapour, plant transpiration or xylem water in the gas phase. 12 , 41 , 42 With the development and application of new water stable isotopologue measurement techniques, 2 , 8 , 10 , 12 , 43 , 44 which are based on the direct measurement of water vapour of different mixing ratios, the water mixing ratio‐dependency of stable water isotopologues with CRDS is gaining relevance. The direct equilibration method to sample, for instance, soil water vapour is based on thermodynamic equilibrium and full water vapour saturation, which is controlled by the ambient temperature that typically varies considerably under natural conditions. This means that a change in soil temperature from 15 to 25°C results in a change of 9000 ppm. Thus, isotopologue values should be corrected to a target value of ∼ 18,000 ppm (default mixing ratio in Picarro software), which requires an instrument‐specific calibration. However, few studies correct or at least report instrument‐specific corrections of the isotopic compositions for different water mixing ratios, 12 , 43 , 44 which might lead to misinterpretations of ecohydrological processes due to an erroneous analysis.

Both, fractionation or contamination during CVE and water mixing ratio dependencies during water stable isotopologue analysis via CRDS might considerably contribute to uncertainties in water stable isotopologue research. Thus, the objectives of this study were to 1) investigate the effect of two different cover materials with and without organics for CVE on δ2H and δ18O values and 2) determine the water mixing ratio dependency of δ2H and δ18O values for water stable isotopologue analysis with CRDS.

2. EXPERIMENTAL

2.1. Sample material, preparation and CVE

To test the effects of cotton wool on δ2H and δ18O values, we chose water samples with a wide range of isotopic compositions (Table 1). These included a variety of in‐house standards, tap water (Freiburg, Germany) and varying precipitation samples from a Mediterranean site (Portugal). 45 , 46 Every sample was divided into four subsamples with ∼ 0.75 mL each. The first subsample was immediately filtered with a glass fibre syringe filter (0.7 µm, Syringe Filter; Membrane Solutions) to remove potential particle contaminations, especially of precipitation samples, and set aside as reference (a). The three other subsamples were frozen (‐20°C) immediately in 1.5 mL glass vials (LLG) and prepared for CVE as in Kübert et al. 2 Freezing sample material (e.g. soil, xylem) is a common practice to avoid the loss of water vapour during the evacuation procedure of the sampling line. To ensure the frozen status of the sample, the vials were additionally inserted in liquid N2 (−196°C) right before extraction. 47 As cover material, we used (b) commercially available cotton wool (elkos FACE; EDEKA), (c) glass wool (extra fine, Karl Hecht GmbH & Co KG) or (d) none at all. CVE was conducted on a custom‐made extraction line designed by R. Siegwolf (similar to Ehleringer & Dawson 3 ), which allowed the extraction of 20 samples simultaneously. Sample lines were evacuated to remove any residual water vapour in the extraction lines before the samples were lowered into the water bath. In the water bath, samples were heated to 95°C for 90 min and connected to U‐tubes immersed in a cold trap filled with liquid N2. During extraction, the whole system was connected to a pump with a vacuum of 0.1 mbar (XDS10 vacuum pump; Edwards). 2 After CVE, U‐tubes containing the frozen water were sealed airtight until samples were completely defrosted. Then, the extracted water was filtered with a glass fibre syringe filter (0.7 µm, Syringe Filter; Membrane Solutions), transferred into 1.5 mL glass vials (LLG) and immediately measured with the reference samples (a) on an A0325 autosampler and a V1102‐i vaporization module operated at 140°C connected to a Picarro 2130‐i CRDS instrument (Picarro Inc.) with N2 as carrier gas. In total, nine injections of each sample were used, of which the first six were discarded to avoid carry‐over effects of the previous sample. All four subsamples (a–d) were measured consecutively in a random order. Care was taken that the water mixing ratio for measurements was very similar between samples at ∼18,000 ppm. For calibration, we used four in‐house water standards (Sample ID 1–4 in Table 1), which are regularly calibrated against the international standards VSMOW‐2, SLAP‐2 and GRESP (IAEA). No organic contamination was detected by ChemCorrect (Picarro Inc.). Hydrogen and oxygen stable isotopic ratios are reported relative to the values of the Vienna Standard Mean Ocean Water and expressed as isotopic composition (respectively, δ2H and δ18O) in per mill (‰). 48

TABLE 1

Water samples were used to test the effect of cotton and glass wool on δ2H and δ18O values during cryogenic vacuum extraction. Samples were (A) only filtered (reference), (B) extracted with cotton wool, (C) extracted with glass wool and (D) extracted without cover material. δ2H (+1 SD) and δ18O (+1 SD) values are given for the sample, which was only filtered (A) and measured directly with a Picarro 2130i laser spectrometer.

Sample IDSample descriptionδ18O (‰)δ2H (‰)
1Standard—Light−25.05 ± 0.01−102.44 ± 0.05
2Standard—Medium−9.23 ± 0.02−63.75 ± 0.04
3Standard—Heavy−5.14 ± 0.03−9.66 ± 0.03
4Standard—Superheavy−0.37 ± 0.0254.32 ± 0.08
5Standard—Light 2−15.15 ± 0.02−111.22 ± 0.03
6Standard—Heavy 22.89 ± 0.001.77 ± 0.05
7Tap water Alps−24.96 ± 0.02−101.99 ± 0.08
8Tap water Black Forest−9.27 ± 0.03−63.58 ± 0.08
9Tap water University of Freiburg−9.37 ± 0.01−64.46 ± 0.13
10Distilled water University of Freiburg−9.34 ± 0.03−63.94 ± 0.12
11Precipitation Vila Viçosa 12 Jan 2018−7.65 ± 0.05−48.23 ± 0.07
12Precipitation Vila Viçosa 12 Jun 2018−2.00 ± 0.04−6.12 ± 0.17
13Precipitation Vila Viçosa 16 May 20180.07 ± 0.018.87 ± 0.08
14Precipitation Vila Viçosa 25 Sep 2019−2.02 ± 0.02−4.97 ± 0.04
15Precipitation Vila Viçosa 24 Oct 2018−7.71 ± 0.03−47.03 ± 0.10
16Precipitation Vila Viçosa 5 Feb 2019−4.72 ± 0.01−24.77 ± 0.05
17Precipitation Vila Viçosa 21 Oct 2019−3.28 ± 0.03−20.20 ± 0.05
18Precipitation Vila Viçosa 3 Dec 2019−5.88 ± 0.03−33.76 ± 0.09
19Precipitation Vila Viçosa 17 Dec 2019−7.80 ± 0.02−46.85 ± 0.06
20Groundwater Vila Viçosa 23 Mar 2020−5.35 ± 0.01−32.08 ± 0.06

2.2. Sample material for water mixing ratio test

To investigate the dependency of δ2H and δ18O on the analysed water mixing ratio in CRDS, we selected four in‐house standards (Sample ID 1–4 in Table 1) that span a wide isotopic range (δ2H: −102.90‰/δ18O: −25.13‰, −64.01‰/−9.28‰, −10.27‰/−5.22‰ and 53.89‰/−0.40‰). We measured with different sample volumes on the Picarro 2120‐i CRDS instrument coupled to an A0325 autosampler and a V1102‐i vaporization module (Picarro Inc.) and dry air as carrier gas (see above). The sample volume can be adjusted in the Picarro program Autosampler Control. The program allows us to run a sequence of measurements with different methods (i.e. in our study sample volumes) automatically. The program's default is 1.8 µL. The nine chosen amounts varied between 0.2 and 3.1 µL to cover a wide range of water concentrations. For each amount and sample, eight injections were conducted, of which the first three were discarded for analysis. This procedure was repeated with a second Picarro 2130‐i CRDS instrument with eight different sample volumes between 0.25 and 3.1 µL to compare the performance of the CRDS models. The results of this analysis are reported and illustrated as uncalibrated raw values to demonstrate the variability of δ2H and δ18O before an instrument‐specific calibration is applied.

2.3. Statistical analysis

We compared δ2H and δ18O values of water samples extracted with cotton, glass wool or without cover material and reference samples with a linear mixed effect model (lme) 49 with treatment (reference—cotton wool—glass wool—without cover) as fixed and sample ID (see Table 1) as a random factor. In case of statistical differences, a post‐hoc pairwise comparison of the means with the function emmeans with a Bonferroni adjustment in the package emmeans 50 in R was applied. 51 As the requirements for a parametric test were not always met, we repeated the analysis with the non‐parametric Friedman Test (friedman_test) in the R package rstatix 52 and the Wilcox Test (wilcox_test) with a Bonferroni adjustment for a pairwise comparison of the different treatments. 51

To visualize the offset of δ2H and δ18O values from the reference samples, we applied a linear mixed effect model (lme) 49 with measured δ2H/δ18O values (reference) and treatment (cotton wool—glass wool—without cover) as fixed and sample ID as a random effect.

In addition, we calculated the deviation of the isotopic composition of samples extracted with cotton wool, glass wool and without cover material from the isotopic composition of the reference sample:

Δδ=δmeasured−δreference

and applied a linear model (lm) between Δδ2H/Δδ18O and δ2H/δ18O of the reference samples to test for increasing/decreasing deviations with more enriched/depleted samples for each type of cover material separately.

For the relationship of δ2H and δ18O values and water mixing ratios in CRDS, we applied non‐linear third‐order logarithmic regression in the form of:

δ=a+b∗lnWMR+c∗(lnWMR)2+d∗lnWMR3,

where δ is the δ2H or δ18O (in‰), WMR is the water mixing ratio (in ppm), and a, b, c and d regression parameters. For consistency, we used the same non‐linear regression for both isotopes and for both CRDS models. We are aware that the application of this regression may result in an overfitting. However, the sole purpose of this regression was to better visualize the trends of δ2H and δ18O values with increasing water mixing ratio and not to provide a water mixing ratio correction matrix. All graphics and the non‐linear third‐order logarithmic regression were conducted with SigmaPlot (Version 14, Systat Software GmbH).

3. RESULTS

3.1. Effect of cover material on δ2H values

CVE with cotton instead of glass wool or without cover had a clear and significant effect on the isotopic composition of the extracted water samples (Figure 1 and Table 2). δ2H values were significantly more depleted (p < 0.001, Figure 1A) for cotton extracted compared to reference samples, with an average deviation of 7.2 ± 4.7‰ (mean ± standard deviation, SD). For example, at δ2H values of +54.3‰ (Sample ID 4, Table 1), cotton‐extracted water samples were depleted by 14.3‰ compared to reference samples (Figures 1A and 2A). For 2H‐depleted samples, such as the light in‐house standard (δ2H = −102.44‰, Sample ID 1), deviations were less pronounced (Figures 1A and 2A). The increasing fractionation effect with more enriched samples for cotton wool extracted samples was significant (p < 0.001, R2 = 0.71, Figure 2A). δ2H values of water extracted with glass wool (average deviation: δ2H: 0.8 ± 0.8‰, δ18O: 0.1 ± 0.1‰, Figure 1C) or without cover material (average deviation: δ2H: 0.8 ± 0.8‰, δ18O: 0.1 ± 0.1‰, Figure 1E) did not differ strongly from reference samples (both p > 0.05, Figures 1 and 2A and Table 2). Interestingly, δ2H values of water extracted with glass wool also deviated stronger from the reference sample with increasing enrichment (Figure 2A). This trend was weak and only marginally significant (p = 0.06, R2 = 0.14). For δ2H values of water extracted without cover, this pattern was not evident (p > 0.05) (Figure 2A).

An external file that holds a picture, illustration, etc. Object name is ANSA-5-2300053-g001.jpg

Relationship of reference and extracted δ2H values, as well as the relationship of reference and extracted δ18O values for cotton wool (A, B), glass wool (C, D) and without cover (E, F) for cryogenic vacuum extraction with 1:1 line (dotted) and linear mixed effect model. All regressions were highly significant (p < 0.001)***.

TABLE 2

Average values with a standard deviation of δ2H and δ18O values for the four different treatments of the 20 water samples: Reference, extracted with cotton, extracted with glass wool and extracted without cover material. Different letters indicate statistical differences at a significance level of p < 0.05 (linear mixed effect model & Friedmann Test) for each isotope separately.

TreatmentReferenceCottonGlass woolWithout cover
δ2H (‰)−39.00 ± 41.23a −45.63 ± 36.64b −39.53 ± 40.81a −38.85 ± 41.16a
δ18O (‰)−7.57 ± 7.25a −7.56 ± 6.89a −7.61 ± 7.21a −7.57 ± 7.22a
An external file that holds a picture, illustration, etc. Object name is ANSA-5-2300053-g002.jpg

Relationship of reference δ2H and offset (Δδ2H) of extracted and reference δ2H value (extracted—reference) (A) and relationship of reference δ18O and offset (Δδ18O) of extracted and reference δ18O value (measured—reference) (B) with linear regression. Significance level (linear regression) given at p < 0.05*, p < 0.01**, p < 0.001***.

3.2. Effect of cover material on δ18O values

For δ18O, the values after extraction with cotton wool, glass wool or without cover did not differ significantly from the reference values (Figure 1B,D,F and Table 2). However, similar trends as for δ2H values were evident for the different cover materials. δ18O values for samples extracted with cotton wool showed the strongest deviation from the reference values (0.3 ± 0.2‰). This deviation was lowest at δ18O values of ∼ −9‰ and significantly increased with more enriched or more depleted samples (p < 0.001, R2 = 0.90, Figure 2B). This trend was also evident in a minor deviation from the 1:1 line in Figure 1B. Samples extracted with glass wool had a low deviation from the reference samples in their isotopic composition (0.1 ± 0.1‰, Figure 1D). However, this deviation significantly increased with more enriched samples (p < 0.05, R2 = 0.21). This trend was, as for δ2H, only weak for δ18O (Figure 2B). Samples extracted without cover material had similar deviations from the reference values in δ18O (0.1 ± 0.1‰, Figure 3C) compared to samples extracted with glass wool, yet no significant trend was evident (p > 0.05). In summary, the trends were similar for both isotopes, 2H and 18O, although they were more significant for 2H. Samples extracted without cover material showed the best agreement with and did not differ significantly from the reference samples (Figures 1 and 2 and Table 2).

An external file that holds a picture, illustration, etc. Object name is ANSA-5-2300053-g003.jpg

Relationship of uncalibrated δ2H and water mixing ratios measured with two Picarro cavity ring‐down spectroscopy (CRDS) instruments (L2120‐i and L2130‐i) for four water samples with different isotopic compositions (Table 1, sample ID 1–4): (A) light, (B) medium, (C) heavy and (D) superheavy. The dotted line represents the true value as calibrated with IAEA standards.

3.3. Effect of water mixing ratio in CRDS on δ2H and δ18O values

For both isotopes, 2H and 18O, the measured water mixing ratio had a strong impact on both evaluated CRDS models (Figures 3 and 4). With increasing water mixing ratios, the CRDS instrument measurements indicated more enriched isotopic compositions compared to the true value (except for δ18O values of the light standard, Figure 4A). This effect levelled off for δ2H values above 10000 ppm compared to the increase below 10000 ppm (Figure 3). For δ18O values, the obtained isotopic composition increased almost linearly after 10000 ppm (Figure 4). Changes in measured isotopic compositions below 10,000 ppm were mostly non‐linear for both isotopes and CRDS instruments. The comparison of the CRDS models clearly indicated that the measurements with the L2130‐i were less variable over the measured water mixing ratio range for both isotopes (Figures 3 and 4). Please note that the values reported are uncalibrated raw values for both instruments.

An external file that holds a picture, illustration, etc. Object name is ANSA-5-2300053-g004.jpg

Relationship of uncalibrated δ18O and water mixing ratios measured with two Picarro cavity ring‐down spectroscopy (CRDS) instruments (L2120‐i and L2130‐i) for four water samples with different isotopic compositions (Table 1, sample ID 1–4): (A) light, (B) medium, (C) heavy and (D) superheavy. The dotted line represents the true value as calibrated with IAEA standards.

4. DISCUSSION

4.1. Cotton wool strongly fractionates during CVE

Cotton wool as cover material had a strong effect on the isotopic composition of the analysed samples. The depleted δ2H values compared to the reference samples are in agreement with other studies testing fractionation effects for cotton 53 or other organic materials. 14 , 32 Here, different mechanisms played a role. As potentially 30% of the hydrogen atoms are exchangeable in cellulose, 54 , 55 such hydrogen exchange with the extracted water vapour could explain the strong fractionation effects for 2H. At higher temperatures, the kinetic energy of molecules increases, which can facilitate isotope exchange. 14 However, isotope exchange with cellulose as the sole explanation for the fractionation effects cannot elucidate the patterns observed for δ18O of samples extracted with cotton wool. A variety of studies concluded that fractionation effects caused by CVE do not affect δ18O values of samples, 13 , 14 , 53 which is in contrast to our results. Indeed, the effect (Figure 2B) for δ18O was small but still significant. Thus, in addition to hydrogen exchange, there seems to be a mechanism affecting both stable isotopologues of water. Here, the hygroscopicity of cotton wool 22 , 26 , 27 might play a decisive role. As we did not dry the cotton wool prior to the extraction, adsorbed atmospheric water vapour was potentially replaced with water vapour originating from the water sample during CVE. Chen et al. found the isotopic composition of water, which was adsorbed by organic surfaces, including cotton, was clearly depleted in 2H compared to unconfined water. 32 They attributed this to “surface effects”, where water adsorbed by large surfaces is isotopically different from bulk water molecules due to structural and energetic differences. 32 Such adsorption and surface effects would also be highly relevant for δ18O and could, thus, explain our results.

As we observed similar patterns for both isotopes of glass wool, although of much lower magnitudes (Figure 2), surface effects might also explain these patterns. Richard et al. demonstrated that surface effects were also present for porous silica tubes as adsorbent, 56 which are chemically similar to glass wool. Nevertheless, deviations of δ2H and δ18O values of water samples extracted with glass wool from reference water samples were very small.

In conclusion, these results illustrate the strong fractionation effect of cotton wool on water isotopic composition. They also demonstrate the urgent need to report if and what kind of cover materials are used in CVE. If possible, any cover material should be avoided. In cases, where a cover material is necessary (i.e. for soil samples), glass wool is an appropriate choice. In addition, test runs of CVE systems with pure water of different isotopic compositions are strongly recommended.

4.2. CVE works well—for pure water samples

While extraction with both, cotton‐ and glass wool, showed (small) fractionation effects, isotopic compositions of samples extracted without cover did not demonstrate any difference from reference samples. This illustrates that CVE is by no means an unsuitable technique as suggested in recent years. 15 It did not produce any significant measurement uncertainties with our protocol. This statement is only valid for the extraction of pure water samples and cannot inform on effects occurring with plant 13 , 14 , 31 or soil 23 , 24 materials (which was not the aim of this study). However, even if fractionation effects do occur, they do not necessarily invalidate study results obtained by using CVE. 22 , 57 On one hand, some experiments reporting fractionation effects were conducted under artificial conditions, such as spiking experiments testing the effect of soil properties on the isotopic composition during CVE. 22 Here, soil samples are artificially oven‐dried at 105°C, and rewetted with water of known isotopic composition, which is then extracted cryogenically. 15 , 23 Such high temperature and fast drying do not represent natural conditions and processes and thus, might lead to different isotopic results. On the other hand, (measurement) uncertainties can rarely be excluded from studies and should be accounted for. 57 Apart from measurement uncertainties, practical considerations should also be reviewed. For example, in‐situ methods are rather ideal for small‐scale, mechanistic studies, 2 , 12 which cannot (yet) cover the spatial heterogeneity of water stable isotopologue composition naturally occurring in ecosystems. Here, destructive sampling and CVE can be, for instance, combined with in‐situ methods to address both the spatial and temporal resolution. 12

δ2H and δ18O values are strongly dependent on the analysed water mixing ratio in CRDS

Our results demonstrate that newly developed methods 8 , 11 to measure the isotopic compositions directly in the water vapour need to correct for water mixing ratio effects and isotope dependencies. 37 Such methods have little control over the water mixing ratio and measure over a large range of natural variations in mixing ratios, so measurements must be made at the given water mixing ratio and cannot easily be adjusted in‐situ to, for example, ∼18,000 ppm, as recommended by Picarro. This is a known problem 37 , 38 and corrections are commonly applied. 11 , 12 , 58 However, there are several issues we would like to point out with this study. From our data, it is evident that water mixing ratios below 10,000/5000 ppm (δ2H/δ18O) yield variable and non‐linear changes in isotopic compositions for both, δ2H and δ18O values in CRDS, which makes corrections more difficult. Thus, low water mixing ratios, although given as an appropriate measurement range in CRDS, 39 should be treated with caution, as they might yield unreliable results. As an alternative, if such low water mixing ratios are unavoidable, such as in cooler environments, the CRDS instrument should be characterised thoroughly for this range. As demonstrated, these effects are clearly instrument‐specific 37 and might even change over time. Therefore, the water mixing ratio dependency of each instrument should be individually determined. We recommend determining this dependency based on at least three water isotope standards, spanning a wide isotopic range that covers the expected measurement range, and in steps of, for example, 0.4 µL. 12 We also recommend making use of the Picarro Autosampler Control and autosampler to run an automated sequence. Alternatively, head space calibration in combination with a dilution of dry air or N2 can be used. We further suggest first correcting the water mixing ratio dependency, for example, to correct all isotopic compositions to, for instance, the target value of ∼ 18,000 ppm (default mixing ratio in Picarro software). The water mixing ratio correction should be based on a water isotope standard in a similar isotopic range to the measured value to be corrected. Then, the newly derived isotopic composition can be calibrated against a calibration curve, determined for the target mixing ratio range (e.g. ∼ 18,000 ppm) and ideally from at least three water standards, spanning a wide isotopic range of the expected measurement values. 12 , 37 If these steps are followed, new in‐situ methods provide an enormous advantage and can provide insights into ecohydrological dynamics and processes on a (sub) daily basis. 42 , 43 , 59

5. CONCLUSION

The usage of the inappropriate cover material for CVE and negligence of different water mixing ratios for CRDS measurements can strongly modify the obtained results for δ2H and δ18O. If not corrected adequately, these results might lead to erroneous or invalid conclusions about (ecohydrological) processes. Thus, we strongly recommend following and reporting strict measurement protocols in water‐stable isotopologue research with CVE and CRDS.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ACKNOWLEDGEMENTS

We would like to acknowledge funding from the DFG (WE 2681/10‐1).

Open access funding enabled and organized by Projekt DEAL.

Notes

Haberstroh S, Kübert A, Werner C. Two common pitfalls in the analysis of water‐stable isotopologues with cryogenic vacuum extraction and cavity ring‐down spectroscopy. Anal Sci Adv. 2024;5:2300053. 10.1002/ansa.202300053 [CrossRef] [Google Scholar]

DATA AVAILABILITY STATEMENT

Data are available from figshare doi:10.6084/m9.figshare.24961440.v1.

REFERENCES

1. Sprenger M, Leistert H, Gimbel K, Weiler M. Illuminating hydrological processes at the soil‐vegetation‐atmosphere interface with water stable isotopes. Rev Geophys. 2016;54(3):674‐704. 10.1002/2015RG000515 [CrossRef] [Google Scholar]
2. Kübert A, Paulus S, Dahlmann A, et al. Water stable isotopes in ecohydrological field research: comparison between in situ and destructive monitoring methods to determine soil water isotopic signatures. Front Plant Sci. 2020;11:387. 10.3389/fpls.2020.00387 [PMC free article] [PubMed] [CrossRef] [Google Scholar]
3. Ehleringer JR, Dawson TE. Water uptake by plants: perspectives from stable isotope composition. Plant Cell Environ. 1992;15(9):1073‐1082. 10.1111/j.1365-3040.1992.tb01657.x [CrossRef] [Google Scholar]
4. Dubbert M, Werner C. Water fluxes mediated by vegetation: emerging isotopic insights at the soil and atmosphere interfaces. New Phytol. 2019;221(4):1754‐1763. 10.1111/nph.15547 [PubMed] [CrossRef] [Google Scholar]
5. Dawson TE, Ehleringer JR. Isotopic enrichment of water in the “woody” tissues of plants: implications for plant water source, water uptake, and other studies which use the stable isotopic composition of cellulose. Geochim Cosmochim Acta. 1993;57(14):3487‐3492. 10.1016/0016-7037(93)90554-A [CrossRef] [Google Scholar]
6. Dalton FN. Plant root water extraction studies using stable isotopes. Plant Soil. 1988;111:217‐222. [Google Scholar]
7. Orlowski N, Frede H‐G, Brüggemann N, Breuer L. Validation and application of a cryogenic vacuum extraction system for soil and plant water extraction for isotope analysis. J Sens Sens Systems. 2013;2(2):179‐193. 10.5194/jsss-2-179-2013 [CrossRef] [Google Scholar]
8. Volkmann THM, Weiler M. Continual in situ monitoring of pore water stable isotopes in the subsurface. Hydrol Earth Syst Sci. 2014;18(5):1819‐1833. 10.5194/hess-18-1819-2014 [CrossRef] [Google Scholar]
9. Beyer M, Kühnhammer K, Dubbert M. In situ measurements of soil and plant water isotopes: a review of approaches, practical considerations and a vision for the future. Hydrol Earth Syst Sci. 2020;24(9):4413‐4440. 10.5194/hess-24-4413-2020 [CrossRef] [Google Scholar]
10. Magh R‐K, Gralher B, Herbstritt B, et al. Technical note: conservative storage of water vapour—practical in situ sampling of stable isotopes in tree stems. Hydrol Earth Syst Sci. 2022;26(13):3573‐3587. 10.5194/hess-26-3573-2022 [CrossRef] [Google Scholar]
11. Volkmann THM, Kühnhammer K, Herbstritt B, Gessler A, Weiler M. A method for in situ monitoring of the isotope composition of tree xylem water using laser spectroscopy. Plant Cell Environ. 2016;39(9):2055‐2063. 10.1111/pce.12725 [PubMed] [CrossRef] [Google Scholar]
12. Kübert A, Dubbert M, Bamberger I, et al. Tracing plant source water dynamics during drought by continuous transpiration measurements: an in‐situ stable isotope approach. Plant Cell Environ. 2022;46(1):133‐149. 10.1111/pce.14475 [PubMed] [CrossRef] [Google Scholar]
13. Chen Y, Helliker BR, Tang X, Li F, Zhou Y, Song X. Stem water cryogenic extraction biases estimation in deuterium isotope composition of plant source water. Proc Natl Acad Sci USA. 2020;117(52):33345‐33350. 10.1073/pnas.2014422117 [PMC free article] [PubMed] [CrossRef] [Google Scholar]
14. Wen M, He D, Li M, Ren R, Jin J, Si B. Causes and factors of cryogenic extraction biases on isotopes of xylem water. Water Resour Res. 2022;58(8). 10.1029/2022WR032182 [CrossRef] [Google Scholar]
15. Orlowski N, Breuer L, Angeli N, et al. Inter‐laboratory comparison of cryogenic water extraction systems for stable isotope analysis of soil water. Hydrol Earth Syst Sci. 2018;22(7):3619‐3637. 10.5194/hess-22-3619-2018 [CrossRef] [Google Scholar]
16. West AG, Patrickson SJ, Ehleringer JR, Water extraction times for plant and soil materials used in stable isotope analysis. Rapid Commun Mass Spectrom. 2006;20(8):1317‐1321. doi: 10.1002/rcm.2456 [PubMed] [CrossRef] [Google Scholar]
17. Jia G, Yu X, Deng W, Liu Y, Li Y. Determination of minimum extraction times for water of plants and soils used in isotopic analysis. J Food Agric Environ. 2012;10:1035‐1040. Accessed: July 21, 2022. https://www.semanticscholar.org/paper/Determination‐of‐minimum‐extraction‐times‐for‐water‐Li‐Deng/2c3c036abdb77d0fa9c78f3cdf83fd645cd0d7df [Google Scholar]
18. Tsuruta K, Yamamoto H, Katsuyama M, Kosugi Y, Okumura M, Matsuo N. Effects of cryogenic vacuum distillation on the stable isotope ratios of soil water. Hydrol Res Lett. 2019;13(1):1‐6. 10.3178/hrl.13.1 [CrossRef] [Google Scholar]
19. Orlowski N, Breuer L, McDonnell JJ. Critical issues with cryogenic extraction of soil water for stable isotope analysis. Ecohydrology. 2016;9(1):1‐5. 10.1002/eco.1722 [CrossRef] [Google Scholar]
20. Gaj M, Kaufhold S, McDonnell JJ. Potential limitation of cryogenic vacuum extractions and spiked experiments. Rapid Commun Mass Spectrom. 2017;31(9):821‐823. 10.1002/rcm.7850 [PubMed] [CrossRef] [Google Scholar]
21. Meißner M, Köhler M, Schwendenmann L, Hölscher D, Dyckmans J. Soil water uptake by trees using water stable isotopes (δ2H and δ18O)−a method test regarding soil moisture, texture and carbonate. Plant Soil. 2014;376(1‐2):327‐335. 10.1007/s11104-013-1970-z [CrossRef] [Google Scholar]
22. Newberry SL, Nelson DB, Kahmen A. Cryogenic vacuum artifacts do not affect plant water‐uptake studies using stable isotope analysis. Ecohydrology. 2017;10(8):e1892. 10.1002/eco.1892 [CrossRef] [Google Scholar]
23. Wen M, Si B, Lu Y, Wang H. Water recovery rate and isotopic signature of cryogenic vacuum extracted spiked soil water following oven‐drying at different temperatures. Hydrol Processes. 2021;35(6). 10.1002/hyp.14248 [CrossRef] [Google Scholar]
24. Araguás‐Araguás L, Rozanski K, Gonfiantini R, Louvat D. Isotope effects accompanying vacuum extraction of soil water for stable isotope analyses. J Hydrol. 1995;168(1‐4):159‐171. 10.1016/0022-1694(94)02636-P [CrossRef] [Google Scholar]
25. Barbeta A, Burlett R, Martín‐Gómez P, et al. Evidence for distinct isotopic compositions of sap and tissue water in tree stems: consequences for plant water source identification. New Phytol. 2022;233(3):1121‐1132. 10.1111/nph.17857 [PubMed] [CrossRef] [Google Scholar]
26. Masson DO, Richards ES. On the hygroscopic action of cotton. Proc R Soc Lond A. 1906;78(525):412‐429. 10.1098/rspa.1906.0089 [CrossRef] [Google Scholar]
27. Plante AM, Holcombe BV, Stephens LG. Fiber hygroscopicity and perceptions of dampness. Text Res J. 1995;65(5):293‐298. 10.1177/004051759506500507 [CrossRef] [Google Scholar]
28. Fang DD, ed. General description of cotton. in: Cotton Fiber: Physics, Chemistry and Biology. Springer; 2018:1‐11. [Google Scholar]
29. McCall ER, Jurgens JF. Chemical composition of cotton. Text Res J. 1951;21(1):19‐21. 10.1177/004051755102100105 [CrossRef] [Google Scholar]
30. Wassenaar LI, Hobson KA. Improved method for determining the stable‐hydrogen isotopic composition (δD) of complex organic materials of environmental interest. Environ Sci Technol. 2000;34(11):2354‐2360. 10.1021/es990804i [CrossRef] [Google Scholar]
31. Diao H, Schuler P, Goldsmith GR, Siegwolf RTW, Saurer M, Lehmann MM. Technical note: on uncertainties in plant water isotopic composition following extraction by cryogenic vacuum distillation. Hydrol Earth Syst Sci. 2022;26(22):5835‐5847. 10.5194/hess-26-5835-2022 [CrossRef] [Google Scholar]
32. Chen G, Auerswald K, Schnyder H. 2H and 18O depletion of water close to organic surfaces. Biogeosciences. 2016;13(10):3175‐3186. 10.5194/bg-13-3175-2016 [CrossRef] [Google Scholar]
33. Schmidt M, Maseyk K, Lett C, et al. Reducing and correcting for contamination of ecosystem water stable isotopes measured by isotope ratio infrared spectroscopy. Rapid Commun Mass Spectrom. 2012;26(2):141‐153. 10.1002/rcm.5317 [PubMed] [CrossRef] [Google Scholar]
34. Johnson JE, Hamann L, Dettman DL, et al. Performance of induction module cavity ring‐down spectroscopy (IM‐CRDS) for measuring δ18 O and δ2 H values of soil, stem, and leaf waters. Rapid Commun Mass Spectrom. 2017;31(6):547‐560. 10.1002/rcm.7813 [PubMed] [CrossRef] [Google Scholar]
35. West AG, Goldsmith GR, Brooks PD, Dawson TE. Discrepancies between isotope ratio infrared spectroscopy and isotope ratio mass spectrometry for the stable isotope analysis of plant and soil waters. Rapid Commun Mass Spectrom. 2010;24(14):1948‐1954. 10.1002/rcm.4597 [PubMed] [CrossRef] [Google Scholar]
36. Millar C, Janzen K, Nehemy MF, Koehler G, Hervé‐Fernández P, McDonnell JJ. Organic contamination detection for isotopic analysis of water by laser spectroscopy. Rapid Commun Mass Spectrom. 2021;35(15):e9118. 10.1002/rcm.9118 [PubMed] [CrossRef] [Google Scholar]
37. Schmidt M, Maseyk K, Lett C, et al. Concentration effects on laser‐based δ18 O and δ2 H measurements and implications for the calibration of vapour measurements with liquid standards. Rapid Commun Mass Spectrom. 2010;24(24):3553‐3561. 10.1002/rcm.4813 [PubMed] [CrossRef] [Google Scholar]
38. Komiya S, Kondo F, Moossen H, et al. Characterizing water vapour concentration dependence of commercial cavity ring‐down spectrometers for continuous on‐site atmospheric water vapour isotope measurements in the tropics. Atmos Meas Tech. 2021;14(2):1439‐1455. 10.5194/amt-14-1439-2021 [CrossRef] [Google Scholar]
39. Inc P. Isotopic water analyzer for δ18O and δD. [Datasheet]. 2021.
40. Gupta P, Noone D, Galewsky J, Sweeney C, Vaughn BH. Demonstration of high‐precision continuous measurements of water vapor isotopologues in laboratory and remote field deployments using wavelength‐scanned cavity ring‐down spectroscopy (WS‐CRDS) technology. Rapid Commun Mass Spectrom. 2009;23(16):2534‐2542. 10.1002/rcm.4100 [PubMed] [CrossRef] [Google Scholar]
41. Dubbert M, Kübert A, Werner C. Impact of leaf traits on temporal dynamics of transpired oxygen isotope signatures and its impact on atmospheric vapor. Front Plant Sci. 2017;8:5. 10.3389/fpls.2017.00005 [PMC free article] [PubMed] [CrossRef] [Google Scholar]
42. Kühnhammer K, van Haren J, Kübert A, et al. Deep roots mitigate drought impacts on tropical trees despite limited quantitative contribution to transpiration. Sci Total Environ. 2023;893:164763. 10.1016/j.scitotenv.2023.164763 [PMC free article] [PubMed] [CrossRef] [Google Scholar]
43. Kinzinger L, Mach J, Haberstroh S, et al. Interaction between beech and spruce trees in temperate forest ecosystem affects water use, root water uptake pattern and canopy structure. Tree Physiol. 2023:tpad144. [PubMed] [Google Scholar]
44. Kühnhammer K, Dahlmann A, Iraheta A, et al. Continuous in situ measurements of water stable isotopes in soils, tree trunk and root xylem: field approval. Rapid Commun Mass Spectrom. 2022;36(5):e9232. 10.1002/rcm.9232 [PubMed] [CrossRef] [Google Scholar]
45. Haberstroh S, Caldeira MC, Lobo‐do‐Vale R, et al. Nonlinear plant‐plant interactions modulate impact of extreme drought and recovery on a Mediterranean ecosystem. New Phytol. 2021;231(5):1784‐1797. 10.1111/nph.17522 [PubMed] [CrossRef] [Google Scholar]
46. Haberstroh S, Lobo‐do‐Vale R, Caldeira MC, Dubbert M, Cuntz M, Werner C. Plant invasion modifies isohydricity in Mediterranean tree species. Funct Ecol. 2022;36:2384‐2398. 10.1111/1365-2435.14126 [CrossRef] [Google Scholar]
47. Koeniger P, Marshall JD, Link T, Mulch A. An inexpensive, fast, and reliable method for vacuum extraction of soil and plant water for stable isotope analyses by mass spectrometry. Rapid Commun Mass Spectrom. 2011;25(20):3041‐3048. 10.1002/rcm.5198 [PubMed] [CrossRef] [Google Scholar]
48. Gonfiantini R. Standards for stable isotope measurements in natural compounds. Nature. 1978;271(5645):534‐536. 10.1038/271534a0 [CrossRef] [Google Scholar]
49. Comprehensive R Archive Network . nlme: linear and nonlinear mixed effects models. 2022. Accessed: July 12, 2023. https://CRAN.R‐project.org/package=nlme
50. Comprehensive R Archive Network . emmeans: estimated marginal means, aka least‐squares means. 2022. Accessed: July 12, 2023. https://CRAN.R‐project.org/package=emmeans
51. R Core Team . R: a language and environment for statistical computing. R Foundation for Statistical Computing.
52. Comprehensive R Archive Network . rstatix: pipe‐friendly framework for basic statistical tests. 2022. Accessed: July 12, 2023. https://CRAN.R‐project.org/package=rstatix
53. Jiang H, Gu H, Chen H, Sun H, Zhang X, Liu X. Comparative cryogenic extraction rehydration experiments reveal isotope fractionation during root water uptake in Gramineae. New Phytol. 2022;236(4):1267‐1280. 10.1111/nph.18423 [PubMed] [CrossRef] [Google Scholar]
54. Epstein S, Yapp CJ, Hall JH. The determination of the D/H ratio of non‐exchangeable hydrogen in cellulose extracted from aquatic and land plants. Earth Planet Sci Lett. 1976;30(2):241‐251. 10.1016/0012-821X(76)90251 [CrossRef] [Google Scholar]
55. Meier‐Augenstein W, Kemp HF, Schenk ER, Almirall JR. Discrimination of unprocessed cotton on the basis of geographic origin using multi‐element stable isotope signatures. Rapid Commun Mass Spectrom. 2014;28(5):545‐552. 10.1002/rcm.6811 [PubMed] [CrossRef] [Google Scholar]
56. Richard T, Mercury L, Massault M, Michelot J‐L. Experimental study of D/H isotopic fractionation factor of water adsorbed on porous silica tubes. Geochim Cosmochim Acta. 2007;71(5):1159‐1169. 10.1016/j.gca.2006.11.028 [CrossRef] [Google Scholar]
57. Allen ST, Kirchner JW. Potential effects of cryogenic extraction biases on plant water source partitioning inferred from xylem‐water isotope ratios. Hydrol Processes. 2022;36(2):e14483. 10.1002/hyp.14483 [CrossRef] [Google Scholar]
58. Dubbert M, Cuntz M, Piayda A, Werner C. Oxygen isotope signatures of transpired water vapor: the role of isotopic non‐steady‐state transpiration under natural conditions. New Phytol. 2014;203(4):1242‐1252. 10.1111/nph.12878 [PubMed] [CrossRef] [Google Scholar]
59. Werner C, Meredith LK, Ladd SN, et al. Ecosystem fluxes during drought and recovery in an experimental forest. Science. 2021;374(6574):1514‐1518. 10.1126/science.abj6789 [PubMed] [CrossRef] [Google Scholar]
Articles from Analytical Science Advances are provided here courtesy of Wiley
-