Solid phase extraction of explosives on Ni-doped carbosils prepared by mechanochemistry

Abstract

The article presents the application of two sets of Ni-doped carbosils in the solid phase extraction of explosives. The adsorbents were prepared by two different methods. The first set of carbosils was obtained by mechanochemical deposition of potato starch and nickel salt on the surface of silica gel, and subsequent carbonization. The second set of carbosils was obtained from the same precursors and under quite similar conditions, i.e. with the exception of mechanochemical deposition of potato starch replaced by the gelation step. The prepared adsorbents were applied in solid phase extraction of explosive nitrate esters, and nitroaromatics from aqueous solutions. The adsorption and desorption steps were evaluated separately. It was found that textural properties, influenced by carbon deposit and nickel content, have a large impact on the solid phase extraction results. The recovery rates obtained onto carbosils prepared by mechanochemical method are approximately thrice as high as those observed for carbosils prepared by gelation method. It was shown that the composites with moderate nickel content can be used as effective materials for extraction both of aliphatic and aromatic explosives.

Introduction

Carbon-inorganic oxides composites, especially based on silica gels or nanosilicas (carbosils), have proved their high potential in many applications (Gogotsi and Presser, 2013). Adsorption properties of carbosils depend on the silica matrix, conditions of pyrolysis and of a carbon precursor (Blitz and Gun’ko, 2006; Hubbard, 2002). The dual surface properties of carbosils are caused by a mosaic-like surface structure, which consists both of a nonpolar carbon deposit and of a polar silica. Therefore, these hybrids can be used for solid phase extraction (SPE) both of organics (Rudziński et al., 1995), and inorganic, e.g. toxic metal ions (Seledets et al., 2005). Previously carbosils proved to be promising materials for SPE of explosives (Tomaszewski and Gun’ko, 2015; Tomaszewski et al., 2015). These preliminary results proved that complex polar/nonpolar surface of carbosils had diversified selectivity towards aliphatic, cyclic or aromatic explosives. Previously two new set of carbosils were prepared and tested (Charmas, 2015a, 2015b). The first set of adsorbents was obtained by mechanochemical treatment, and subsequent carbonization (Charmas, 2015b). The second set of adsorbents was obtained from the same precursors and under quite similar conditions, i.e. with the exception of mechanochemical treatment replaced by the gelation step (Charmas, 2015a). The prepared Ni–doped materials had unique micro/mesoporous structure. Therefore, the aim of this work was to evaluate the influence of the surface characteristics of Ni-doped carbosils on the SPE performance with respect to explosives. The preliminary data concerning these experiments were presented quite recently (Tomaszewski et al., 2016).

Experimental analysis

Preparation of carbon–silica adsorbents (carbosils)

Carbosils were prepared on the basis of porous silica gel Sipernat 50 (S-50, Degussa, Germany), potato starch (Superior, Poland) as a carbon source and nickel(II) nitrate hexahydrate (POCh, Poland). The initial carbosil without nickel and those doped with different amounts of metal were prepared. The molar ratio of silica/starch/Ni²⁺ was 1/0.3/0.0–2.5. The homogeneous suspension of nickel salt, starch and silica in water was heated at 69℃. After 1 h, the starch underwent gelation. The prepared mixtures were air-dried, disintegrated and sieved to obtain 0.16–0.2 mm grain fraction. These materials were pyrolysed in a rotary reactor in a nitrogen flow at 500℃ for 3 h. The carbosils were denoted as GCS-Ni-x, where x = 0.0, 0.5,…, 2.5 is the relative nickel content. The mechanochemical step of the preparation of the second set of carbosils was carried out in the planetary mill PULVERISETTE 7 (Fritsch, Germany) according to the procedure: pre-grinding (15 min, 500 rpm) and final grinding (30 min, 750 rpm). Additionally, to examine the influence of a mechanochemical treatment on porous structure of silica, it was subjected to the same treatment (adsorbent S-50 M). All the silica/starch/nickel salt mixtures were subjected to the described mechanochemical treatment. Then the samples were pyrolysed as above. The samples obtained in this way are denoted with the general formulae MCS-Ni-x, where x = 0.0, 0.5,…, 2.5 is the relative nickel content.

Methods

Nitrogen adsorption

The nitrogen adsorption/desorption isotherms were recorded using a Micromeritics ASAP 2405N adsorption analyser. The specific surface area S_BET was calculated using the standard BET equation (Gregg and Sing, 1982). The total pore volume V_p was estimated under the relative pressure p/p₀ 0.98. The nitrogen desorption data were used to calculate the pore-size distribution (PSD) using the procedure described elsewhere (Gun’ko and Mikhalovsky, 2004). The volumes and specific surface area of micropores, mesopores and macropores were calculated as previously (Charmas, 2015a).

SPE

Three millilitres of disposable syringes containing 100 mg of carbosil were rinsed with 5 ml of acetonitrile (ACN) and conditioned with 5 ml of water. The 10 ml aqueous samples (containing 5 vol.% of ACN) were applied; the effluents were collected. The 5 ml ACN extracts collected into 10 ml flasks were made up with water. The initial concentration of explosives in samples was 5 µg/ml. The following explosives were used: nitrate esters: triethyleneglycol dinitrate (TEGDN), trinitroglycerin (NG), pentaerythritol tetranitrate (PETN) and nitroaromatics: 2,4-dinitrotoluene (DNT), 2,4,6-trinitrotoluene (TNT) and 1,3,5-trinitrobenzene (TNB). Using the results of HPLC quantification, the parameters describing the SPE results were calculated: adsorption rate R_ads, desorption rate R_des and recovery rate R_rec (Tomaszewski et al., 2015).

Results and discussion

Nitrogen adsorption

Table 1 presents the porous structure characteristics of GCS and MCS carbosils, Figure 1 shows their nitrogen adsorption/desorption isotherms and Figure 2 the PSD curves. The detailed analysis of structural and textural properties of the studied carbosils was presented previously (Charmas, 2015a, 2015b), therefore in this article only the most important characteristics are discussed. Taking into consideration the IUPAC recommendation, the isotherms observed for both series of carbosils belong to the IV type, but the hysteresis loops belong to the type H3. This course of isotherms indicates the microporous structure of the adsorbents containing slit-like pores and cylindrical mesopores. Data from Table 1 indicate that the MCS samples are characterized by higher S_BET, S_micro and S_meso, in comparison to GCS ones. This correlates well with the shift of MCS isotherm branches towards higher values of nitrogen adsorption (Figure 1). For both series of carbosils, the increase of the nickel content did not bring about systematic changes in S_BET. However, at the same time, a systematic decrease of S_micro and increase of S_meso was observed. On the other hand, the differences in the contribution of micropores (V_micro/V_p) and mesopores (V_micro/V_p) for most of the studied carbosils are not significant. Namely the contribution of micropores is slightly higher for GCS samples, while the contribution of mesopores is somewhat higher for MCS ones.

Table 1.

Textural characteristics of initial silicas and Ni-dopped carbosils prepared by mechanochemical or gelation method.

Sample	SBET (m2/g)	Smicro (m2/g)	Smeso (m2/g)	Smacro (m2/g)	Vp (cm3/g)	Vmicro (cm3/g)	Vmeso (cm3/g)	Vmacro (cm3/g)	Vmicro/Vp (%)	Vmeso/Vp (%)
Sipernat 50	359	74	247	38	1.583	0.024	0.840	0.719	2	53
Sipernat 50 M	232	35	180	8	0.583	0.012	0.441	0.130	2	76
MCS-Ni-0.0	188	121	65	2	0.171	0.046	0.073	0.052	27	43
MCS-Ni-0.5	232	130	100	2	0.194	0.048	0.109	0.037	25	56
MCS-Ni-1.0	243	126	115	2	0.228	0.045	0.140	0.043	20	61
MCS-Ni-1.5	250	102	147	1	0.245	0.035	0.183	0.027	14	75
MCS-Ni-2.0	224	74	149	1	0.259	0.026	0.214	0.019	10	83
MCS-Ni-2.5	246	79	166	1	0.315	0.027	0.259	0.029	9	82
GCS-Ni-0.0	167	107	60	–	0.149	0.040	0.086	0.023	27	58
GCS-Ni-0.5	127	90	37	–	0.089	0.036	0.045	0.008	40	51
GCS-Ni-1.0	118	77	40	1	0.135	0.033	0.063	0.039	24	47
GCS-Ni-1.5	115	76	39	–	0.095	0.029	0.059	0.007	31	62
GCS-Ni-2.0	118	64	52	2	0.163	0.024	0.103	0.036	15	63
GCS-Ni-2.5	113	50	62	1	0.442	0.046	0.321	0.072	10	73

Note: S_BET is the specific surface area; S_micro, S_meso and S_macro are the specific surface area of micro-, meso- and macropores; V_p is the pore volume; V_micro, V_meso and V_macro are the volume of micro, meso and macropores.

Figure 1.

Nitrogen adsorption/desorption isotherms for carbosils prepared by gelation method (left graph) and by mechanochemical method (right graph).

Figure 2.

Pore-size distributions for carbosils prepared by gelation method (left graph) and by mechanochemical method (right graph).

SPE

The adsorption rates of explosives onto MCS and GCS adsorbents are presented in Table 2. The R_ads values on MCS for nitrate esters and nitroaromatics averages 65% and 93%, respectively. A growth of the Ni content initially results in an increase in R_ads of nitrate esters on MCS. The highest values for TEGDN and NG were obtained for MCS-Ni-1.0 sample, thereafter the subsequent increase of the Ni content results in lowering of R_ads. A quite similar changes of R_ads were observed for PETN but the highest R_ads was obtained for MCS-Ni-1.5 sample. The changes of the adsorption effectiveness of nitrate esters can be attributed to the properties of the carbon deposit. On the one hand, an increase of the mesoporosity (and decrease of the micropores content) promotes the adsorption of aliphatic explosives (Tomaszewski and Gun’ko, 2015), on the other hand, the increase of the Ni amount decrease the content of carbon deposit and consequently the number of available adsorption sites. Also the increase of the doping metal content can support the formation of carbides (Nagy et al., 2014) or accelerate the graphitization of carbon deposit (Xia et al., 2012). Such hydrophobic surfaces, i.e. carbides and graphene planes in particular (Leboda et al., 2001) can weaken the adsorption of aliphatic explosives. The R_ads for nitroaromatics onto MCS are substantially higher than that for nitrate esters. This comes from the stronger adsorption affinity of nitroaromatics, as compared to the nitrate esters, resulting from π–π interactions between the nitroaromatics and the graphene planes of carbon deposits (Tomaszewski and Gun’ko, 2015). The adsorption rates of nitroaromatics are similar for all of the MCS samples and are not depended on the Ni content. R_ads on GCS adsorbents are smaller and in some cases accounts for as little as half (for nitroaromatics), or even one-third (for nitrate esters) of the values attained for MCS ones. The observation stay in a good agreement with the courses of the nitrogen adsorption/desorption isotherms (Figure 1), i.e. lower nitrogen adsorption on GCS samples, as well lower S_BET, S_micro or S_meso (Table 1). Values of R_ads both for nitrate esters and for nitroaromatics are independent on the Ni content in GCS adsorbents.

Table 2.

Adsorption and desorption rates of explosives on carbosils prepared by mechanochemical or gelation method.

Adsorbent	Adsorption rate Rads (%)						Desorption rate Rdes (%)
	DNT	TNT	TNB	TEGDN	NG	PETN	DNT	TNT	TNB	TEGDN	NG	PETN
Mechanochemical method
MCS-Ni-0.0	93	84	80	35	36	62	69	70	69	58	76	85
MCS-Ni-0.5	97	98	94	70	87	86	87	85	85	89	91	82
MCS-Ni-1.0	96	95	98	80	89	80	90	86	90	92	86	98
MCS-Ni-1.5	96	97	98	63	76	97	95	95	97	86	86	88
MCS-Ni-2.0	97	95	96	46	61	78	95	90	95	77	70	91
MCS-Ni-2.5	93	89	81	30	35	52	93	90	93	50	64	97
Gelation method
GCS-Ni-0.0	43	44	49	25	29	42	74	70	81	60	65	65
GCS-Ni-0.5	41	48	44	20	30	46	75	59	75	72	64	70
GCS-Ni-1.0	42	45	43	18	24	40	70	54	70	61	55	60
GCS-Ni-1.5	46	47	42	16	26	39	77	55	72	76	59	61
GCS-Ni-2.0	47	53	44	13	25	48	76	62	69	77	62	66
GCS-Ni-2.5	45	52	41	21	28	40	86	61	85	50	67	71

The desorption rates of explosives from MCS and GCS adsorbents are presented in Table 2. The R_des obtained on MCS for nitrate esters and nitroaromatics are quite similar and averages 81% and 87%, respectively. For the MCS and nitrate esters, a growth of the nickel content influences the R_des values. It is similar to the changes observed for R_ads, i.e. maximum of R_des can be observed. By an example for TEGDN and PETN, the highest R_des values was obtained on MCS-Ni-1.0 adsorbent, while for NG on MCS-0.5 Ni one. Also for nitroaromatics, R_des is dependent on the Ni content. The maximum of R_des was observed but for the adsorbent of a higher Ni content, i.e. MCS-Ni-1.5. At this stage of discussion, it is difficult to explain why the quite similar R_des was achieved for aliphatic and aromatic explosives. In this context, it is important to recall that R_ads for nitroaromatic were substantially higher than for nitrate esters. Therefore, one could suspect an opposite effect during desorption step, i.e. higher R_des of esters and lower of nitroaromatic (Tomaszewski and Gun’ko, 2015). It should be borne in mind that this effect can be attributed to the different interactions of aliphatic and aromatic molecules with carbonaceous patches of carbosil (Tomaszewski and Gun’ko, 2015). The less effective adsorption of nitrate esters, especially NG or TEGDN, might result from the tendency of small nitroaliphatic molecules to form dimers (Sagawa and Shikata, 2014), characterized by substantially lower adsorption. The average R_des of explosives on GCS are ca. 15% smaller than the values for MCS ones. The probable reason is a porosity of GCS adsorbents that are characterized by a more complex pore shape and higher relative contribution of micropores (Table 1). Therefore, desorption from narrower pores is less effective due to the strong interactions with pore walls.

The most important parameter describing the efficiency of the SPE method is the recovery rate (R_rec), which depends both on the adsorption and desorption rates (Tomaszewski and Gun’ko, 2015). It determines the effectiveness of analyte isolation from samples, and thus the selectivity of the SPE method. R_rec both for nitrate esters and for nitroaromatics on MCS are approximately thrice as high as those observed for GCS (Figure 3). The low recoveries obtained on GCS, i.e. average for esters 19% and for nitroaromatics 32% arises, first of all, from the very ineffective adsorption step. Also the uncompleted desorption of explosives makes negative contribution to the final value of R_rec. The data (Figure 3) indicate that the best adsorbent for extraction of nitrate esters from aqueous samples is carbosil MCS-Ni-1.0, with the recoveries reaching nearly 80%. For nitroaromatic compounds adsorbent MCS-Ni-1.5 proved its high performance, namely recoveries reaching 95% were obtained.

Figure 3.

Dependence of the SPE recovery rates (R_rec) of explosives on relative nickel content in carbosils. Left graph for nitrate esters: squares for TEGDN, circles for NG and triangles for PETN; right graph for nitroaromatics: squares for TNB, circles for DNT and triangles for TNT. Full symbols for MCS-Ni-x samples, empty ones for GCS-Ni-x ones. The lines are polynomial fittings of the R_rec values, added to guide eyes on the changes of R_rec.

Conclusion

Two series of the hybrid carbon/silica adsorbents was synthesized by the carbonization of the mixture of starch and silica gel Sipernat 50 doped with nickel salt. The initial mixtures containing the same quantity of the reagents were homogenized by two different methods, i.e. by mechanochemical treatment and by manual mixing followed by gelation step. The application of mechanochemistry, despite the destruction of the initial structure of porous silica, results in the larger specific surface area and larger pore volume of the prepared in that way carbosils. It has been also observed that the nickel doping causes increase in the mesoporosity and decrease in the microporosity for both series of adsorbents. The carbosils prepared by the mechanochemical method proved to be very useful as adsorbents in the SPE of explosive nitrate esters and nitroaromatics from aqueous samples. The results – SPE recovery rates are 2–3 times larger in comparison to the results for carbosils prepared without mechanochemical treatment. The better SPE performance of the mechanochemically prepared carbosils in the first place may results from the homogenous application of carbon precursor (starch) and metal catalyst (nickel) on the surface of silica matrix, what causes the uniformly distributed carbon deposit, as well from its unique porous structure and the shape of adsorbent particles.