Small and wide angle X-ray studies of impregnated activated carbons

CARBON

7 5 ( 2 0 1 4 ) 4 2 0 –4 3 1

Available at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/carbon

Small and wide angle X-ray studies of impregnated activated carbons J.W.H. Smith a, M. Mcdonald a, J.V. Romero a, L. MacDonald a, J.R. Lee a, J.R. Dahn a b

a,b,*

Department of Physics and Atmospheric Science, Dalhousie University, Halifax, Nova Scotia B3H 3J5, Canada Department of Chemistry, Dalhousie University, Halifax, Nova Scotia B3H 4J3, Canada

A R T I C L E I N F O

A B S T R A C T

Article history:

Small angle X-ray scattering (SAXS) and powder X-ray diffraction (XRD) measurements

Received 13 January 2014

were made on CuO-impregnated activated carbons, prepared with and without an HNO3

Accepted 5 April 2014

co-impregnant, in order to determine the effect of impregnant loading and HNO3 content

Available online 13 April 2014

on impregnant distribution. A comprehensive matrix of 30 CuO-impregnated samples with five HNO3 concentrations and six impregnant loadings was prepared and studied. As a highlight, in Cu-based samples prepared with no HNO3, relatively small particle size CuO impregnant (approximately 3 nm) was observed at low impregnant loading and additional CuO appeared in large particles (>10 nm diameter) in meso and macropores as the impregnant loading increased. By contrast, when 4 M HNO3 was present during the impregnation, the largest impregnant particles found were less than 4 nm. Results from SAXS data were shown to be in good agreement with XRD and data obtained from nitrogen gas adsorption isotherms. The combination of SAXS and XRD is shown to be a powerful combination in elucidating the nanostructure of impregnated activated carbons.  2014 Elsevier Ltd. All rights reserved.

1.

Introduction

Activated carbons (AC) are commonly used in air filtration devices, such as gas mask canisters, to remove organic vapours from contaminated airstreams [1,2]. Impregnating the AC substrate with properly selected chemicals can increase adsorption of certain gases [3,4] or give the ability to adsorb toxins that un-impregnated AC cannot [5–16]. Impregnated activated carbons (IACs) capable of broad spectrum gas adsorption have been reported in the literature [7–11,14,17–19]. Copper oxide (CuO) is a common impregnant in broad spectrum IACs [5,6,8,10,14,17,18]. Our group has reported on the relationship between better impregnant distribution on the AC substrate and better gas adsorption capacity of the IAC [3,16,20–22]. It was observed that

co-impregnating with nitric acid (HNO3) improved the distribution of impregnant on the AC substrate [8,20–22]. In these works, wide angle powder X-ray diffraction (XRD) was found to be a useful tool to examine the distribution of impregnant on the AC substrate for relatively large grain size impregnant (>3–4 nm approximately). The XRD method could not provide information about relatively small nanocrystallites (<3 nm approximately). Small angle X-ray scattering (SAXS) has been demonstrated to be a useful technique to study the structure of AC and other porous carbonaceous materials [23–30]. SAXS methods have also been employed to study two phase inorganic systems [31] and the particle size of supported catalysts [32,33]. In a recent report [34] a method for analyzing the SAXS data obtained from IACs was demonstrated. In the work

* Corresponding author at: Department of Physics and Atmospheric Science, Dalhousie University, Halifax, Nova Scotia B3H 3J5, Canada. Fax: +1 902 494 5191. E-mail address: [email protected] (J.R. Dahn). http://dx.doi.org/10.1016/j.carbon.2014.04.021 0008-6223/ 2014 Elsevier Ltd. All rights reserved.

CARBON

reported here a comparative study of some factors influencing impregnant distribution on IACs prepared from aqueous copper nitrate Cu(NO3)2 (Cu(NO3)2) solutions is reported. Some of the IACs were co-impregnated with nitric acid (HNO3). SAXS and XRD were used to study the changes in the IACs as a function of Cu(NO3)2 and HNO3 concentration. Experimental results obtained using the SAXS and XRD methods were compared and found to be in good agreement. Experimental results from SAXS measurements were also compared to data obtained from nitrogen (N2) gas adsorption isotherms and good agreement was observed.

2.

Experimental details

2.1.

Chemicals used

The chemicals used to prepare the impregnating solutions were reagent grade copper nitrate hemi-pentahydrate (Cu(NO3)22.5H2O) and 70% concentrated HNO3 obtained from Sigma–Aldrich.

with argon flowing at approximately 200 mL/min. The argon flow rate was approximately 60 mL/min during heating. The IACs were heated at TF = 180 C for approximately 3 h. After heating all of the IACs were cooled under flowing argon until they reached room temperature. The impregnant loading after heating was determined using the equation: % Loading ¼ fðmassfinal  massinitial Þ=massinitial g  100%;

ð1Þ

where massfinal is the mass of the IAC after heating, and massinital is the mass of the unimpregnated GC (determined after drying at 120 C in air). The 15 g IAC samples were weighed using a Sartorious balance with an accuracy of ±1 mg. The uncertainty associated with the % loading is 2–3% primarily due to inability to avoid some exposure to moist air while weighing and storing the samples. The predicted impregnant loading (expressed in mmol impregnant/ g GC) was estimated from the volume and concentration of impregnating solution added to the GC. The relative uncertainty is 0.1 mmol impregnant/g GC.

2.3. 2.2.

421

7 5 (2 0 1 4) 4 2 0–43 1

Sample characterization

Sample preparation

All of the IAC samples were prepared using Kuraray GC granular activated carbon (Kuraray Chemical Co. (Osaka, Japan)). Kuraray GC (referred to as ‘‘GC’’ from this point forward) is derived from coconut shells and has a particle mesh size of 12 · 35 (particle size of 1.70 mm · 0.5 mm) and an ash content of 0.4% (wt/wt). Nitrogen adsorption experiments were performed using a Micromeritics ASAP 2010 gas adsorption porosimeter (described below). The data was analyzed using the Brunauer–Emmet–Teller (BET) model (for specific surface area) and non-local density functional theory (for pore size distribution (PSD)) on software supplied by Micromeritics. Details of these measurements have been reported previously [35]. Table 1 summarizes the results of the gas adsorption measurements performed on GC carbon. Boehm titrations [36] have been performed on GC in previous work [37,38] and show that this carbon has a total of 60.18 mmol/g acidic surface groups. IAC samples were impregnated with Cu(NO3)2 and in some cases HNO3. The concentrations of Cu(NO3)2 used were 0.31 M, 0.79 M, 1.60 M, 2.40 M and 3.10 M and the concentrations of HNO3 used were 0 M, 0.5 M, 2.0 M, 4.0 M and 8.0 M. Due to solubility issues, the IAC prepared from the most concentrated acid solution was 2.0 M Cu(NO3)2/6.6 M HNO3. All of the IACs were impregnated using the imbibing or incipient wetness method [37,38]. Typically 11–12 mL of solution was added to approximately 15 g of GC. The IAC samples were heated under flowing argon in a sealed, cylindrical aluminum container that was located inside an oven. Prior to heating, the container was purged

Contact angle measurements were performed using a First Ten Angstroms (FTA) 135 drop shape analyzer. A drop of solution was formed on the tip of a syringe. The drop was deposited onto a highly oriented pyrolytic graphite (HOPG) substrate by slowly lowering the syringe until the drop contacted the substrate. Then the syringe was lifted up. The base of the deposited drop was approximately 3 mm in diameter. A snapshot of the drop was then taken and the data was imported into a computer where the drop shape was analyzed using software supplied by the manufacturer. The HOPG substrates were obtained from SPI Supplies. Grades of SPI-1 and SPI-3 were used and there were no discernible differences in the results obtained from the two different grades. The HOPG was cleaved after each contact angle measurement using adhesive tape. The reported contact angle measurements are an average of 6–12 measurements per solution. X-ray diffraction experiments were performed using a Phillips PW 1720 X-ray generator operated at a voltage of 40 kV and a current of 30 mA. The system is equipped with a Cu target X-ray tube and an incident beam monochromator which selects Cu Ka radiation. The system is coupled to an Inel CPS 120 curved position sensitive detector. A typical dwell time used was 1200 s/sample. Samples were ground to a fine powder prior to measurement using a mortar and pestle. SAXS experiments were performed using a Bruker Nanostar system. The X-ray generator operates at 40 kV and 650 lA. The system is equipped with a Cu target X-ray tube and has Go¨bel mirrors and a series of pinhole collimators to select a 400 lm diameter beam of Cu Ka radiation

Table 1 – Characteristics calculated from N2 adsorption experiments performed on GC. BET surface area (m2/g)

Micropore volume (cm3/g)

Meso and micropore volume (cm3/g)

1550

0.53

0.55

422

CARBON

7 5 ( 2 0 1 4 ) 4 2 0 –4 3 1

˚ ). This system is equipped with a two (wavelength, k = 1.5418 A dimensional area detector (Vantec-2000, Bruker AXS). A typical experiment has a scattering angle range from 0.23 to 5.0 ˚ 1) and a counting time of (approximately 0.016–0.356 A 1000 s/sample. Samples for SAXS experiments were finely ground with a mortar and pestle and packed in a 0.16 cm thick stainless steel sample holder that had 9 rectangular holes. The volume of each hole was approximately 0.03 cm3. Scotch MagicTM tape (3 M Co.) was used on both sides of the hole to hold the sample in place. The un-impregnated GC samples had a mass of approximately 12 ± 1 mg and a thickness of 1.62 ± 0.02 mm. The impregnated samples had the same thickness, but higher mass, due to the presence of the impregnant. Additional details of the SAXS experiments are included in the Supplementary material and have been reported earlier [34]. Nitrogen adsorption experiments were performed using a Micromeritics ASAP 2010 porisimeter. The data was corrected for the free-space in the sample tube and for any deviations from the ideal gas law. The amount of N2 adsorbed per gram of adsorbent is reported at a temperature and pressure of 273.15 K and 1 atm, respectively. Approximately 0.5 g of AC or IAC sample (pre-dried at 120 C) was added to the sample tube. Samples were evacuated at an elevated temperature until any adsorbed species, such as water, were removed. When dealing with IAC samples the temperature applied during degassing was always less than the maximum final heating temperature (TF) the IAC was exposed to during sample preparation. Degassing continued until out-gassing from the sample was negligible, at least 24–36 h for the samples examined in this work. Measurements were made up to a relative pressure (P/Po) of 0.9. Desorption measurements were also performed by applying a vacuum to the system’s manifold until a lower partial pressure was achieved and then allowing the system to reach equilibrium with the valve to the sample tube open.

Fig. 1 – Contact angle measurements plotted against Cu(NO3)2 concentration for Cu(NO3)2 and Cu(NO3)2/HNO3 solutions. The concentration of HNO3 is indicated in the legend. For reference, the measured contact angle for distilled water has been included. The lines are guides for the eye. (A colour version of this figure can be viewed online.)

3.

Results and discussion

concentration. The Cu(NO3)2/HNO3 solutions follow a similar trend; however the contact angle for these solutions is always lower than the solutions with no HNO3 present. For example the Cu(NO3)2/4 M HNO3 solutions have contact angles that are approximately 30 lower than those measured for the Cu(NO3)2 (with no HNO3) solutions. Fig. 1 shows that increasing HNO3 concentration in the impregnating solution decreases contact angle. This implies that solutions with higher HNO3 concentration wet the carbon surface better, giving better impregnant distribution.

3.1.

Contact angle measurements

3.2.

Contact angle studies were performed on an HOPG substrate. This type of substrate has been shown to be reasonable substitute for AC when examining liquid–solid interactions in earlier work [37,38]. Fig. 1 shows contact angle data plotted against Cu(NO3)2 concentration in the impregnating solution for Cu(NO3)2 solutions and Cu(NO3)2/HNO3 solutions. The concentration of HNO3 is indicated in the legend. For reference, the measured contact angle for distilled water has been included. Each data point in Fig. 1 is an average of 6–12 measurements and the error bars represent the standard deviation. Fig. 1 shows that the contact angle for distilled water is approximately 81 ± 4, in reasonable agreement with the literature [39] and previous measurements by our group [37,38]. The data from the Cu(NO3)2 solution (with no HNO3) shows a slightly decreasing contact angle with increasing

Impregnant loading

Fig. 2 shows the impregnant loading, after heating at 180 C under argon, for IACs prepared from Cu(NO3)2 and HNO3 solutions. The data is presented as percent impregnant loading as a function of Cu(NO3)2 concentration in the impregnating solution. The concentration of HNO3 used in each sample set is indicated in the legend. The predicted impregnant loading has been indicated for reference. The predicted loading assumes conversion of the Cu(NO3)2 precursor to CuO. The predicted impregnant loading does not account for mass gain due to the introduction of surface oxygen groups from the action of the HNO3 on the AC substrate [40], or for mass loss due to consumption of carbon from the AC substrate that may be evolved as CO or CO2 according to reactions in Ref. [14]. C, H, N and O gas analysis reported in Ref. [22] showed that IACs prepared from Cu(NO3)2 and HNO3 that were heated at 180 C under argon had 60.8 ± 0.3% nitrogen (N) content (wt.). The

CARBON

7 5 (2 0 1 4) 4 2 0–43 1

impregnant grain size (Lhkl) was calculated using the Scherrer equation [41]. The Lhkl values were determined by analyzing the diffraction peaks at approximately 2h = 35.5 and 38.8 (Miller indices (1 1 1) and (1 1 1), respectively) and are available in Table 1 of the Supplemental material. The average integrated area under the CuO diffraction peaks was also calculated and will be presented later below.

% loading (wt.)

20

15

3.4.

10

0.5 M HNO3

5

2 M HNO3 4 M HNO3 8 M HNO3 Predicted CuO loading

0

1

2

3

[Cu(NO3)2] in impregnating solution / M Fig. 2 – Observed impregnant loading (after heating at 180 C under argon) as a function of Cu(NO3)2 concentration in the impregnating solution for samples prepared from 0 M, 0.5 M, 2.0 M, 4.0 M and 8.0 M HNO3. The predicted loading has also been indicated. (A colour version of this figure can be viewed online.)

IACs had similar hydrogen (H) content. GC that was impregnated with 4 M HNO3 then heated at 180 C had approximately 4% (wt.) oxygen content. Fig. 2 shows a reasonably linear increase in impregnant loading with increasing Cu(NO3)2 concentration for all of the IAC samples prepared in this work. The observed loadings are in reasonable agreement with the Cu(NO3)2 precursor decomposing to CuO during heating. At the two lowest Cu(NO3)2 concentrations, the loading increases with HNO3 molarity. Differences between the observed and predicted impregnant loading may be due to the presence of oxygen containing surface groups [22] or small amounts of preadsorbed moisture from the air (see Section 2.2).

3.3.

Small angle X-ray model and fitting

The model used to fit the SAXS data in this work has been discussed and justified in detail in an earlier report [34]. The form of the model used here is a quasi three phase model that describes scattering from the 3 phase system (pores, carbon and impregnant): 2 3

no HNO3

0

423

Wide angle X-ray characterization

Fig. 3 shows wide angle X-ray diffraction (XRD) data collected from all of the samples displayed in Fig. 2. The Bragg peak positions for CuO (JCPDS file number [00-089-2529]) are indicated in each panel for reference. Fig. 3a–c shows that CuO is the dominant impregnant species present on the IACs after heating. Fig. 3a shows increasing intensity of the CuO diffraction peaks as the impregnant loading is increased. As the concentration of HNO3 is increased (panels a ! e) the intensity of the CuO peaks decreases, which indicates that co-impregnation with HNO3 results in smaller, better distributed impregnant. These results are consistent with previous work [22]. The samples described by Fig. 3d (4 M HNO3 co-impregnant) and Fig. 3e (8 M HNO3 co-impregnant) do not have obvious impregnant-related diffraction peaks, even at the highest loadings reported here. The average

q2x Bgr A Cmi 6 7 IðqÞ ¼ I0K 4q2av n þ þ ðq  xqx Þ2  2 5; q 2 2 ð1 þ a2 q2 Þ2 1þb q

ð2Þ

  h where: q is the magnitude of the scattering vector 4p sin , qav2 k is the squared average electron density of large mesopores, macropores and large impregnant grains, found to a first approximation by weighting impregnant electron density and carbon electron density by mass fraction in the sample, qx2 and q2 are the squared electron densities of the impregnant and carbon (graphite), respectively (assuming electron density of the pores is zero), x is the ratio of the volume of impregnant in the micropore to the volume of the unoccupied micropore (Vx/Vmi). The A term is related to the surface area of meso-, macropores and large impregnant grains, the Bgr and Cmi terms are used to calculate the volume of small grain size impregnant and micropores, respectively [34]. The a and b terms are the Debye correlation lengths for small grain size impregnant (<3 nm approximately) and micropores, respectively. The Debye lengths can be related to the radius of gyration (Rg) by: pffiffiffi  pffiffiffi  ð3Þ Rg ¼ 6b; or 6a : Rg is a measure of the average micropore (or particle) size and can be described as the calculated root mean square distance of the scatterers from their center of mass [42]. The n term is a fitting constant that is related to the fractal dimension of the carbon surface [34,43,44]. The pre-factor I0K has been described in detail in earlier work [34]. The model presented in Eq. (2) assumes that micropore size was not significantly changed by the HNO3 treatment. This assumption is consistent with results reported in the literature [43]. Experimental evidence is provided in Fig. 4 of the Supplementary material. The terms A, n, Bgr, a, Cmi and b in Eq. (2) were used as fitting parameters and were optimized by performing a least squares fitting routine (minimized v2) between the calculated and observed data over the entire experimental range. For IAC samples the ‘b’ term was held constant [34]. In this work v2 is a measure of the goodness of fit and is defined as: v2 ¼

N X 2 ½lnðyi Þ  lnðf ðxi ÞÞ ;

ð4Þ

i¼1

where N is the number of data points, f(xi) is the fit to the data at the ith data point and yi is the ith data point.

424

CARBON

7 5 ( 2 0 1 4 ) 4 2 0 –4 3 1

no HNO3 CuO peaks

0.5 M HNO3 CuO peaks

CuO 21 wt.%

CuO 21 wt.%

CuO 16 wt.%

CuO 16 wt.%

CuO 11 wt.%

CuO 11 wt.%

CuO 6 wt.%

CuO 6 wt.%

CuO 2 wt.%

CuO 2 wt.%

GC base carbon

(b)

(a) 2 M HNO3 CuO peaks

Intensity / (counts) (arb. scale)

GC base carbon

4 M HNO3 CuO peaks

CuO 20 wt.%

CuO 19 wt.%

CuO 16 wt.%

CuO 16 wt.%

CuO 12 wt.%

CuO 12 wt.%

CuO 7 wt.%

CuO 8 wt.%

CuO 3 wt.%

CuO 5 wt.% GC base carbon

GC base carbon

(c)

(d) 20

8 M HNO3 CuO peaks

40

60

Scattering angle / 2θ CuO 13 wt.%

CuO 12 wt.%

CuO 7 wt.%

CuO 5 wt.% GC base carbon

(e) 20

40

60

Scattering angle / 2θ Fig. 3 – XRD data obtained from IACs co-impregnated with aqueous Cu(NO3)2 and HNO3. The impregnant loading after heating is indicated in each panel. Panels a, b, c, d and e show data from IACs co-impregnated with 0, 0.5, 2, 4 and 8 M HNO3, respectively. The Bragg peak positions for CuO are indicated for reference. (A colour version of this figure can be viewed online.)

3.5.

Small angle X-ray characterization

Fig. 4 shows SAXS data collected from all of the samples prepared using different concentrations of Cu(NO3)2/HNO3. All the samples shown in Fig. 4 were heated at 180 C, under flowing argon, prior to measurements. The impregnant loading and concentration of HNO3 used are indicated in each panel. The data is represented by symbols. For clarity only every eighth data point is shown. Fits to the data are represented by lines. Parameters from the fits are listed in Table 2.

˚ 1) Fig. 4 shows that the intensity at low q (q 6 0.06 A increases with increasing CuO impregnant loading. This occurs because the impregnant is forming large grains in the macropores and larger mesopores. As the HNO3 concentration in the impregnating solution was increased (panels a ! e) the spread of intensities in each set of samples ˚ 1 the difference in the intendecreased at low q. At q = 0.02 A sities between the GC base carbon data and the data from the samples with the highest CuO loading for different HNO3 concentrations is compared. Specifically, the difference in

CARBON

GC base carbon CuO / 0.5 M HNO3 - 2 wt.%

GC base carbon CuO - 2 wt.% CuO - 6 wt.% CuO - 11 wt.% CuO - 16 wt.% CuO - 21 wt.%

1000000

425

7 5 (2 0 1 4) 4 2 0–43 1

CuO / 0.5 M HNO3 - 6 wt.% CuO / 0.5 M HNO3 - 11 wt.% CuO / 0.5 M HNO3 - 16 wt.% CuO / 0.5 M HNO3 - 21 wt.%

100000

10000

Intensity / (counts / 1000s)

1000

(b)

(a)

1000000

GC base carbon CuO / 2 M HNO3 - 3 wt.%

GC base carbon CuO / 4 M HNO3 - 5 wt.%

CuO / 2 M HNO3 - 7 wt.%

CuO / 4 M HNO3 - 8 wt.%

CuO / 2 M HNO3 - 12 wt.%

CuO / 4 M HNO3 - 12 wt.%

CuO / 2 M HNO3 - 16 wt.%

CuO / 4 M HNO3 - 16 wt.%

CuO / 2 M HNO3 - 20 wt.%

CuO / 4 M HNO3 - 19 wt.%

100000

10000

1000

(c)

(d) 0.1

GC base carbon CuO / 8 M HNO3 - 5 wt.%

1000000

-1

q / (Å )

CuO / 8 M HNO3 - 7 wt.% CuO / 8 M HNO3 - 12 wt.% CuO / 6.6 M HNO3 - 13 wt.%

100000

10000

1000 0.01

(e) 0.1

q / (Å-1) Fig. 4 – SAXS data obtained from IACs co-impregnated with aqueous Cu(NO3)2 and HNO3. The impregnant loading after heating is indicated in each legend. Panels a, b, c, d and e show data from IACs co-impregnated with 0, 0.5, 2, 4 and 8 M HNO3, respectively. (A colour version of this figure can be viewed online.)

intensities in Fig. 4d (the IACs co-impregnated with 4 M HNO3) is 85% less than in Fig. 4a (the IACs with no co-impregnated HNO3). The difference in intensities in Fig. 4e (the IACs coimpregnated with 6.6 M HNO3) is 90% less than in Fig. 4a. This implies that the HNO3 co-impregnation helps to prevent the formation of large impregnant grains. ˚ 1 show similarities for all The data in Fig. 4a–e for q > 0.2 A the samples. In this region, the intensity of the IACs is close to that of the base carbon. This suggests that little impregnant enters the micropores or that the decrease in intensity due to micropore filling is balanced by the increased intensity of scattering from small grain size impregnant particles occupying the meso- and macropores. This will be discussed further below. The data in the plateau region of Fig. 4a–e ˚ 1) shows dramatic differences, especially (q  0.06–0.2 A

between panels a and d. In panel a the intensity in this region initially increases with increasing impregnant loading, however at the highest loading, the SAXS curve intersects the curves obtained from samples with lower CuO loading. The decrease in scattering intensity from small impregnant ˚ 1) and subsequent increase in scattering from (q  0.06–0.2 A ˚ 1) for the samples with large impregnant particles (q 6 0.05 A high CuO loading (and no co-impregnated HNO3) is indicative of impregnant agglomeration on the GC substrate, with clusters of impregnant contributing to the intensity of the SAXS signal at low values of q. As the concentration of co-impregnated HNO3 was increased, the intensity in the plateau region increased with increasing CuO loading. Fig. 4d shows increasing intensity with increasing impregnant loading in the range ˚ 1. This implies an increase in the ˚ 1 6 q 6 0.2 A from 0.06 A

426 Table 2 – Parameters from fits to the SAXS data shown in Fig. 4. Values reported for the Kuraray GC sample are average values extracted from fits to all of the runs shown in Fig. 4. The uncertainty on the GC data represents the deviation from the average measurement.

2 6 11 16 21 4 7 12 16 20 4 7 12 16 20 5 8 12 16 19 5 7 12 13

0 0 0 0 0 0.5 0.5 0.5 0.5 0.5 2 2 2 2 2 4 4 4 4 4 8 8 8 6.6

0.22 0.25 0.37 0.87 0.51 0.22 0.28 0.40 0.99 1.54 0.25 0.26 0.33 0.52 1.01 0.22 0.32 0.31 0.40 0.49 0.16 0.23 0.30 0.35

Cmi · 1022 (cm6/g) 2.18 ± 0.06 Cmi · 1022 (cm6/g)

˚) Rg – GC (A 6.1 ± 0.2 ˚) Rg1 (A

Bgr · 1023 (cm6/g) n/a Bgr · 1023 (cm6/g)

Vx/Vmi n/a Vx/Vmi

˚) Rg – CuO (A n/a ˚) Rg – CuO (A

v2 1.6 ± 0.4 v2

3.6 3.7 3.7 3.7 3.9 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.6 3.5 3.5 3.5

1.94 1.77 1.52 1.66 1.72 1.86 1.79 1.76 1.70 1.65 1.67 1.54 1.53 1.48 1.46 1.66 1.68 1.54 1.50 1.49 1.56 1.56 1.43 1.39

6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1

1.03 3.50 5.45 5.65 3.81 0.78 3.57 7.88 5.96 3.77 0.44 2.04 6.62 8.19 9.23 0.09 1.69 5.35 10.3 13.9 0.00 0.73 4.31 6.76

0.04 0.11 0.23 0.14 0.11 0.06 0.07 0.08 0.10 0.12 0.05 0.09 0.08 0.11 0.13 0.06 0.04 0.10 0.13 0.13 0.05 0.04 0.10 0.13

11.5 11.5 11.5 11.5 11.5 11.5 16.3 17.9 14.0 11.5 11.5 11.5 15.1 14.5 14.1 11.5 11.5 12.6 13.8 14.6 11.5 11.5 11.5 11.9

1.3 0.9 0.3 0.1 0.7 0.9 0.5 0.3 0.1 0.5 1.4 0.8 0.5 0.3 0.4 1.7 1.2 0.7 0.5 0.4 1.9 2.3 1.0 0.7

7 5 ( 2 0 1 4 ) 4 2 0 –4 3 1

GC [HNO3] (M)

N 3.5 ± 0.1 N

CARBON

GC % Loading impregnant (wt.)

I0Kq2A (g/cm6) 0.22 ± 0.04 I0Kqav2A (g/cm6)

CARBON

amount of relatively small (P3 nm) impregnant particles (as detailed in Table 2).

3.6.

427

7 5 (2 0 1 4) 4 2 0–43 1

Nitrogen adsorption isotherms

Nitrogen adsorption isotherms were collected from certain IAC samples to allow for comparisons with the SAXS data. Fig. 5 shows N2 adsorption isotherms collected from IACs in this study. Panel a shows IACs with increasing CuO loading (and no HNO3) and panel b shows IACs with approximately the same CuO loading, with different concentrations of coimpregnated HNO3. Fig. 5a shows that the volume of N2 adsorbed decreases with increasing CuO loading. The one exception to this trend was the IAC with 11 wt.% loading, which had lower N2 adsorption than the sample with 16 wt.% loading. Fig. 5b shows that the volume of N2 adsorbed decreases slightly with increasing

HNO3 concentration. The sample with 1.5 mmol CuO and 6.6 M co-impregnated HNO3 had higher N2 adsorption than the IAC with 1.9 mmol CuO and 4 M HNO3. This was likely due to differences in the CuO loading. The data in Fig. 5 can be classified as type I isotherms [45] and are typical for mainly microporous materials. The small hysteresis loop in these samples can be attributed to the presence of mesopores.

3.7.

Comparative discussion

Analysis of the data in Figs. 3 and 4 was performed to compare results obtained from the IAC samples using XRD and SAXS. Fig. 6a shows the observed SAXS intensity at ˚ 1 plotted against the impregnant loading (mmol q = 0.02 A CuO/g GC) for each series of samples co-impregnated with HNO3. The HNO3 concentrations are indicated in the legend. The data point for GC is the average value calculated from

Vol. N2 adsorbed / g IAC (cm3/ g) at STP

500

400

300

200

GC base carbon 1.8 mmol CuO no HNO3

GC base carbon CuO - 2 wt. % CuO - 6 wt. % CuO - 11 wt. % CuO - 16 wt. % CuO - 21 wt. %

100

(a)

1.8 mmol CuO 0.5 M HNO3 1.8 mmol CuO 2 M HNO3

(b)

1.9 mmol CuO 4 M HNO3 1.5 mmol CuO 6.6 M HNO3

0 0

0.2

0.4

0.6

0.8

0

0.2

0.4

P / Po

0.6

0.8

P / Po

I(0.02 Å-1) / (counts / 1000s)

2000000

Integrated peak area / (counts x deg.)

Fig. 5 – Volume of N2 adsorbed per gram of IAC for samples with increasing CuO loading (panel a) and approximately constant CuO loading with increasing HNO3 concentration (panel b). Details of the samples are provided in the appropriate legend. For reference data from un-impregnated GC has been included in each panel. (A colour version of this figure can be viewed online.)

GC base carbon no HNO3 0.5 M HNO3 2 M HNO3 4 M HNO3 8 M HNO3

1600000

1200000

800000

(a)

400000

0

0

0.5

1

1.5

mmol CuO / g GC

2

2.5

no HNO3 0.5 M HNO3 2 M HNO3 4 M HNO3 8 M HNO3

4000

3000

2000

(b)

1000

0 0

0.5

1

1.5

2

2.5

mmol CuO / g GC

Fig. 6 – Data obtained from IACs prepared with Cu(NO3)2/HNO3 that were heated at 180 C. Panel a shows the observed ˚ 1. Panel b shows the average integrated area of the CuO impregnant peaks located intensity from the SAXS data at q = 0.02 A at scattering angles 2h = 35.5 and 38.8. The data is plotted against impregnant loading. The lines are a guide for the eye. (A colour version of this figure can be viewed online.)

428

CARBON

7 5 ( 2 0 1 4 ) 4 2 0 –4 3 1

0.3 0.03

0.5 M HNO3

no HNO3

0.1

0.01

(b)

Micrograin volume / (cm3 / g)

(a)

0.3 0

0 0.03

4 M HNO3

2 M HNO3

0.2

0.02

Micropore filling fraction

0.2

0.02

0.1

0.01

(d)

(c) 0 0.03

0

0.5

8 M HNO3

1

1.5

2

0 2.5

mmol CuO / g GC 0.2

0.02

0.1

0.01

(e) 0

0

0.5

1

1.5

2

0 2.5

mmol CuO / g GC Fig. 7 – Migrograin volume (black circles) and micropore filling fraction (blue diamonds) plotted versus impregnant loading for the IACs studied in this work. The concentration of co-impregnated HNO3 is indicated in the appropriate legend. The lines are a guide for the eye. (A colour version of this figure can be viewed online.)

the SAXS data obtained from GC (shown in Fig. 4 and Table 2) and the error bars denote the deviation from the average value. Fig. 6b shows the average integrated area under the CuO Bragg peaks located at approximately 2h = 35.5 and 38.8 (Miller indices (1 1 1) and (1 1 1), respectively). The data is plotted against the impregnant loading. The area under the diffraction peaks was integrated using Fityk software [46]. Fig. 6 shows that there is excellent agreement between the observed trends in the SAXS and XRD data regarding the formation of large CuO grains. The trend of increasing amounts of relatively large impregnant (P10 nm approximately) with increased loading is illustrated in the data obtained from IACs prepared without any HNO3 (blue diamonds) in panels a and b. Fig. 6a and b shows that as the concentration of HNO3

present in the impregnating solution is increased, the scattering intensity and hence total volume of large size CuO impregnant grains decreased. The XRD data provides useful information about impregnant grain size and distribution when diffraction peaks are present, however it does not give information about relatively ˚ approximately). To learn how small impregnant grains (630 A the HNO3 co-impregnation is affecting relatively small ˚ 1 range impregnant grains, the SAXS data in the q > 0.06 A is useful. Fig. 7 shows the volume of small grain size impregnant (micrograin volume) and the micropore filling fraction (Vx/Vmi) plotted against impregnant loading for the IACs in this study as detailed in the figure caption. The volume of small grain size impregnant was calculated from the Bgr

Micropore volume (cm3 / g)

CARBON

429

7 5 (2 0 1 4) 4 2 0–43 1

0.6

0.4

0.2 SAXS data N2 adsorption data

(a)

(b)

SAXS data N2 adsorption data SAXS GC base carbon N2 adsorption GC base carbon

0 0

1

2

mmol CuO / g GC

0

2

4

6

HNO3 concentration / M

Fig. 8 – Micropore volume plotted versus impregnant loading for IACs with increasing CuO loading (panel a) and IACS with 1.8 ± 0.1 mmol CuO/g GC impregnant loading and increasing HNO3 concentration plotted versus HNO3 concentration (panel b). The blue diamonds and black circles represent values calculated from fits to the SAXS and N2 adsorption data, respectively. The blue and open black squares in panel b represent SAXS and N2 adsorption values for un-impregnated GC, respectively. (A colour version of this figure can be viewed online.)

values in Table 2 using the methods outlined in Refs. [34] and [47]. The error bars denote the difference between the fitted parameters and the maximum theoretical value (calculated from the volume and concentration of impregnant imbibed in the GC carbon and the bulk density of CuO). Fig. 7a shows that the micrograin volume increases with increasing impregnant loading up to about 1.2 mmol CuO/g GC. After this, the values decrease. The data in Fig. 7a shows that at the highest impregnant loadings there is a decrease in the scattering contribution from small grain size impregnant. This implies that at higher CuO loadings, the impregnant grains are agglomerating and forming larger particles, which contribute to the SAXS scattering signal from larger particle size impregnant. The micropore filling fraction data follows a similar trend, implying that at the highest CuO loadings less impregnant is forming in the micropores relative to samples with lower CuO loading. Fig. 7b–e shows that the micrograin volume increases with increasing impregnant loading. As the concentration of co-impregnated HNO3 increased the micrograin volume increased at each impregnant loading. This implies that co-impregnating with more concentrated HNO3 helps to suppress the formation of large particle size impregnant. Fig. 7 shows that impregnant is entering the micropores at all impregnant loadings and HNO3 concentrations. In panels 7b–e there is a trend of increasing micropore filling with increasing impregnant loading. Comparing this data to the experimental curves in Fig. 4 shows that the decrease in intensity at high q values, expected when micropores fill, is not observed due to the increased scattering intensity from the small particle size impregnant occupying meso- and macropores. Fig. 8 shows micropore volume values calculated from fits to the SAXS and N2 adsorption data. Panel a shows data from IACS with increasing CuO loading and no co-impregnated HNO3. Panel b shows data from IACs with 1.8 ± 0.1 mmol CuO/g GC impregnant loading and increasing amounts of

co-impregnated HNO3. The model used to calculate the micropore volume from the SAXS data assumes the micropores are identical uniform spheres. The micropore volumes were calculated from the Cmi values in Table 2 according to the methods outlined in the literature [34,47]. The micropore volumes from the N2 adsorption experiments were calculated using software supplied by the porisimeter manufacturer (Micromeritics). The calculations used density functional theory and assumed a slit shaped pore [48]. Fig. 8 shows that there is good qualitative agreement between the trends in the micropore data from the SAXS and N2 adsorption experiments. Fig. 8a shows that the micropore volume decreases with increasing CuO loading, which is consistent with some impregnant occupying the micropores. The good agreement between trends in the SAXS and N2 adsorption data lends confidence to our interpretation of the SAXS data and also implies that pore blockage was not significant in these IACs. It is reasonable that the SAXS values are larger than the corresponding gas adsorption values as the SAXS signal can detect micropores smaller than the diameter of the probe molecule (N2) used in the gas adsorption experiments and can also detect closed pores. The magnitude of the calculated values may also be affected by the choice of pore shape used in the fitting model, however choosing a different pore shape for the SAXS model would simply scale the values up or down and would not influence the observed trends in the data. Fig. 8b shows that there is a relatively small decrease in micropore volume with increasing HNO3 concentration. The trends from the SAXS and N2 adsorption data are in good agreement implying that pore blockage was not a major issue in these IACs. For IACs to be capable of both physical and chemical gas adsorption it is important to minimize micropore filling and pore blockage by the impregnant. Fig. 8 shows that for all the IACs studied the majority of micropores were unoccupied and available for adsorption.

430

4.

CARBON

7 5 ( 2 0 1 4 ) 4 2 0 –4 3 1

Conclusions

Contact angle measurements demonstrated that as HNO3 concentration increased, the impregnating solutions contact angle decreased, implying better surface coverage for Cu(NO3)2 solutions co-impregnated with more concentrated HNO3. SAXS and XRD measurements were found to be useful tools for studying impregnant dispersion on AC. Good agreement between the SAXS and XRD methods was observed. It was found that the SAXS method provides information about ˚ approximately) on the relatively small size impregnant (630 A AC substrate. The XRD method could not provide information on such small impregnant nanocrystallites. Co-impregnating the AC substrate with P4 M HNO3, followed by heating at 180 C resulted in relatively small, well dispersed CuO impregnant. IACs co-impregnated with higher concentration HNO3 were observed to have larger volumes of small grain size impregnant. Trends in the micropore volumes calculated from SAXS and N2 gas adsorption measurements were in good agreement. The data implied that pore filling and/or pore blockage by impregnant was not significant in the IACs in this study. It takes approximately 4–5 h to prepare and measure 9 samples using SAXS. It takes 24–48 h to prepare and measure a single sample using N2 gas adsorption porosimetry. The SAXS method was proven to be well-suited to rapid screening of IAC samples in this work. Using a simple spreadsheet model reasonable fits to the SAXS data have been presented. Parameters extracted from these fits have been used to help explain how the impregnant populates the activated carbon.

Acknowledgements The authors would like to thank 3M Canada and NSERC for their financial support. The authors thank Lisa Croll, Simon Smith and Larry Brey for sharing valuable insight and experience in the field of activated carbon research.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.carbon. 2014.04.021.

R E F E R E N C E S

[1] Wood GO. Activated carbon adsorption capacities for vapours. Carbon 1992;30(4):593–9. [2] Wood GO, Moyer ES. A review of the Wheeler equation and comparison of its applications to organic vapor respirator cartridge breakthrough data. Am Ind Hyg Assoc J 1989;50:400–7. [3] Fortier H, Zelenietz C, Dahn TR, Westreich P, Stevens DA, Dahn JR. SO2 adsorption capacity of K2CO3-impregnated activated carbon as a function of K2CO3 content loaded by soaking and incipient wetness. Appl Surf Sci 2007;253(6):3201–7.

[4] Tseng H-H, Wey M-Y. Study of SO2 adsorption and thermal regeneration over activated carbon-supported copper oxide catalysts. Carbon 2004;42(11):2269–78. [5] Alder JF, Fielden PR, Smith SJ. The adsorption of hydrogen cyanide by impregnated activated carbon cloth. Part I: Studies on cobalt and nickel acetates as impregnants for hydrogen cyanide removal. Carbon 1988;26(5):701–11. [6] Alves BR, Clark AJ. An examination of the products formed on reaction of hydrogen cyanide and cyanogen with copper, chromium (6+) and copper-chromium (6+) impregnated activated carbons. Carbon 1986;24(3):287–94. [7] Brey LA, Smith SJ, Weagle GE. Broad spectrum filter system including tungsten-based impregnant and being useful for filtering contaminants from air or other gases. US Patent No. 7309513, 2007. [8] Croll LM, Dahn JR, Siedle AR, Smith JWH, Westreich P, Wood TE. Ammonia-free gaseous air filter. US Patent Application No. 20100050583, 2009. [9] Croll LM, Dahn JR, Smith JWH, Romero JV. Zinc oxide containing filter media and methods of forming the same. US Serial Number 12/941,280. 2010. [10] Doughty DT, Knebel WJ, Cobes JW III. Chromium-free impregnated activated carbon for adsorption of toxic gases and/or vapours in industrial applications. US Patent No. 5492882, 1996. [11] Lodewyckx P. In: Bandosz TJ, editor. Activated carbon surfaces in environmental remediation, vol. 7. Oxford: Elsevier Ltd; 2006. p. 475–523. [12] Nickolov RN, Mehandjiev DR. Comparative study on removal efficiency of impregnated carbons for hydrogen cyanide vapors in air depending on their phase composition and porous textures. J Colloid Interface Sci 2004;273(1):87–94. [13] Barnes PA, Chinn MJ, Dawson EA, Norman PR. Preparation, characterisation and application of metal-doped carbons for hydrogen cyanide removal. Adsorpt Sci Technol 2002;20(9):817–33. [14] Summary Technical Report of Division 10, NDRC, vol. 1, Part 1, Military Problems with Aerosols and Nonpersistent Gases, Washington, DC, 1946 (Available from NTIS, Springfield, VA 22161). [15] Petit C, Karwacki C, Peterson G, Bandosz TJ. Interactions of ammonia with the surface of microporous carbon impregnated with transition metal chlorides. J Phys Chem C 2007;111(34):12705–14. [16] Fortier H, Westreich P, Selig S, Zelenietz C, Dahn JR. Ammonia, cyclohexane, nitrogen and water adsorption capacities of an activated carbon impregnated with increasing amounts of ZnCl2, and designed to chemisorb gaseous NH3 from an air stream. J Colloid Interface Sci 2008;320(2):423–35. [17] Wilson RE, Whetzel JC. Impregnated carbon and process of making same. US Patent No. 1519470, 1921. [18] Smith JWH, Westreich P, Croll LM, Reynolds JH, Dahn JR. Understanding the role of each ingredient in a basic copper carbonate based impregnation recipe for respirator carbons. J Colloid Interface Sci 2009;337(2):313–21. [19] Prasad GK, Singh B, Vijayaraghavan R. Respiratory protection against chemical and biological warfare agents. Defence Sci J 2008;58(5):686–97. [20] Smith JWH, Westreich P, Smith AJ, Fortier H, Croll LM, Reynolds JH, et al. Investigation of copper oxide impregnants prepared from various precursors for respirator carbons. J Colloid Interface Sci 2010;341(1):162–70. [21] Smith JWH, Westreich P, Abdellatif H, Filbee-Dexter P, Smith AJ, Wood TE, et al. The investigation of copper-based impregnated activated carbons prepared from water-soluble materials for broad spectrum respirator applications. J Hazard Mater 2010;180(1–3):419–28.

CARBON

7 5 (2 0 1 4) 4 2 0–43 1

[22] Smith JWH, Romero JV, Dahn TR, Dunphy K, Sullivan B, Mallay M, et al. The effect of heating temperature and nitric acid treatments on the performance of Cu- and Zn-based broad spectrum respirator carbons. J Colloid Interface Sci 2011;364(1):178–94. [23] Kalliat M, Kwak CY, Schmidt PW. In: Blaustein BD, Bockrath BD, Friedman S, editors. New approaches in coal chemistry. Washington DC: American Chemical Society; 1981. p. 3–21. [24] Dubnin MM, Plavnik GM. Microporous structures of carbonaceous adsorbents. Carbon 1968;6(2):183–92. [25] Dahn JR, Xing W, Gao Y. The, ‘‘falling cards model’’ for the structure of microporous carbons. Carbon 1997;35(6):825–30. [26] Yang R, Dahn TR, Dahn JR. Fe–N–C oxygen-reduction catalysts supported on ‘‘burned-off’’ activated carbon. J Electrochem Soc 2009;156(4):B493–8. [27] Stevens DA, Dahn JR. An in situ small-angle X-ray scattering study of sodium insertion into a nanoporous carbon anode material within an operating electrochemical cell. J Electrochem Soc 2000;147(12):4428–31. [28] Gibaud A, Xue JS, Dahn JR. A small angle X-ray scattering study of carbons made from pyrolyzed sugar. Carbon 1996;34(4):499–503. [29] Dash R, Chmiola J, Yushin G, Gogotsi Y, Laudisio G, Singer J, et al. Titanium carbide derived nanoporous carbon for energy-related applications. Carbon 2006;44(12):2489–97. [30] Laudisio G, Dash RK, Singer JP, Yushin G, Gogotsi Y, Fischer JE. Carbide-derived carbons: a comparative study of porosity based on small-angle scattering and adsorption isotherms. Langmuir 2006;22(21):8945–50. [31] Chavez Panduro EA, Beuvier T, Martı´nez MF, Hassani L, Calvigac B, Boury F, Gibaud A. Small-angle X-ray scattering analysis of porous powders of CaCO3. J Appl Crystallogr 2012;45(5):881–5. [32] Stevens DA, Zhang S, Chen Z, Dahn JR. On the determination of platinum particle size in carbon supported platinum electrocatalysts for fuel cell applications. Carbon 2003;41(14):2769–77. [33] De´Ka´ny I, Turi L, Homonnay Z, Ve´rtes A, Burger K. Preparation of nanosize FeS particles on SiO2 and clay mineral supports: SAXS and Mo¨ssbauer spectroscopic measurements. Colloids Surf A: Physicochem Eng Aspects 1996;119(2-3):195–203.

431

[34] McDonald MJ, Smith JWH, Dahn JR. A study of small angle Xray scattering from impregnated activated carbons. Carbon 2014;68:452–61. [35] Smith JWH. A study of respirator carbons [Ph.D. thesis]. Halifax, NS, Canada, Dalhousie University; 2012. [36] Boehm HP. Some aspects of the surface chemistry of carbon blacks and other carbons. Carbon 1994;32(5):759–69. [37] Fortier H. The science of impregnation and the optimization of the performance of impregnated activated carbons for gas mask applications [Ph.D. thesis]. Halifax, NS, Canada, Dalhousie University; 2007. [38] Westreich P, Fortier H, Flynn S, Foster S, Dahn JR. Exclusion of salt solutions from activated carbon pores and the relationship to contact angle on graphite. J Phys Chem C 2007;111(9):3680–4. [39] Adamson AW. Physical chemistry of surfaces. 5th ed. New York: John Wiley & Sons; 1990. p. 397–98. [40] Xiao B, Thomas KM. Adsorption of aqueous metal ions on oxygen and nitrogen functionalized nanoporous activated carbons. Langmuir 2005;21(9):3892–902. [41] Klug HP, Alexander LE. X-ray diffraction procedures: for polycrystalline and amorphous materials. 2nd ed. London: John Wiley & Sons; 1974. p. 687–89. [42] Guinier A, Fournet G. Small-angle scattering of X-rays. New York: John Wiley & Sons; 1955. p. 24–28. [43] La´szlo´ K, Czakkel O, Josepovits K, Rochas C, Geissler E. Influence of surface chemistry on the SAXS response of polymer-based activated carbons. Langmuir 2005;21(18):8443–51. [44] Bale HD, Schmidt PW. Small-angle X-ray-scattering investigation of submicroscopic porosity with fractal properties. Phys Rev Lett 1984;53(6):596–9. [45] Inagaki M. Tascon JMD. In: Bandosz TJ, editor. Activated carbon surfaces in environmental remediation, vol. 7. Oxford; Elsevier Ltd; 2006. p. 55–61. [46] Wojdyr M. Fityk: a general-purpose peak fitting program. J Appl Crystallogr 2010;43:1126–8. [47] Kalliat M, Kwak CY, Schmidt PW. Small angle X-ray measurment of porosity in wood following pyrolosis. Wood Sci Technol 1983;17:241–57. [48] Olivier JP. Modeling physical adsorption on porous and nonporous solids using density functional theory. J Porous Mater 1995;2:9–17.